A balanced and holistic approach to business analytics Business Analytics teaches the fundamental concepts of modern b
807 152 41MB
English Pages 704 [705] Year 2020
Table of contents :
Cover
Half Title Page
Title Page
Copyright Page
Brief Contents
Contents
Preface
About the Author
Credits
Part 1: Foundations of Business Analytics
Chapter 1: Introduction to Business Analytics
Learning Objectives
What Is Business Analytics?
Using Business Analytics
Impacts and Challenges
Evolution of Business Analytics
Analytic Foundations
Modern Business Analytics
Software Support and Spreadsheet Technology
Analytics in Practice: Social Media Analytics
Descriptive, Predictive, and Prescriptive Analytics
Analytics in Practice: Analytics in the Home Lending and Mortgage Industry
Data for Business Analytics
Big Data
Data Reliability and Validity
Models in Business Analytics
Descriptive Models
Predictive Models
Prescriptive Models
Model Assumptions
Uncertainty and Risk
Problem Solving with Analytics
Recognizing a Problem
Defining the Problem
Structuring the Problem
Analyzing the Problem
Interpreting Results and Making a Decision
Implementing the Solution
Analytics in Practice: Developing Effective Analytical Tools at Hewlett‐Packard
Key Terms
Chapter 1 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Appendix A1: Basic Excel Skills
Excel Formulas and Addressing
Copying Formulas
Useful Excel Tips
Excel Functions
Basic Excel Functions
Functions for Specific Applications
Insert Function
Date and Time Functions
Miscellaneous Excel Functions and Tools
Range Names
VALUE Function
Paste Special
Concatenation
Error Values
Problems and Exercises
Chapter 2: Database Analytics
Learning Objectives
Data Sets and Databases
Using Range Names in Databases
Analytics in Practice: Using Big Data to Monitor Water Usage in Cary, North Carolina
Data Queries: Tables, Sorting, and Filtering
Sorting Data in Excel
Pareto Analysis
Filtering Data
Database Functions
Analytics in Practice: Discovering the Value of Database Analytics at Allders International
Logical Functions
Lookup Functions for Database Queries
Excel Template Design
Data Validation Tools
Form Controls
PivotTables
PivotTable Customization
Slicers
Key Terms
Chapter 2 Technology Help
Problems and Exercises
Case: People’s Choice Bank
Case: Drout Advertising Research Project
Part 2: Descriptive Analytics
Chapter 3: Data Visualization
Learning Objectives
The Value of Data Visualization
Tools and Software for Data Visualization
Analytics in Practice: Data Visualization for the New York City Police Department’s Domain Awareness System
Creating Charts in Microsoft Excel
Column and Bar Charts
Data Label and Data Table Chart Options
Line Charts
Pie Charts
Area Charts
Scatter Charts and Orbit Charts
Bubble Charts
Combination Charts
Radar Charts
Stock Charts
Charts from PivotTables
Geographic Data
Other Excel Data Visualization Tools
Data Bars
Color Scales
Icon Sets
Sparklines
Dashboards
Analytics in Practice: Driving Business Transformation with IBM Business Analytics
Key Terms
Chapter 3 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Appendix A3: Additional Tools for Data Visualization
Hierarchy Charts
Waterfall Charts
PivotCharts
Tableau
Problems and Exercises
Chapter 4: Descriptive Statistics
Learning Objectives
Analytics in Practice: Applications of Statistics in Health care
Metrics and Data Classification
Frequency Distributions and Histograms
Frequency Distributions for Categorical Data
Relative Frequency Distributions
Frequency Distributions for Numerical Data
Grouped Frequency Distributions
Cumulative Relative Frequency Distributions
Constructing Frequency Distributions Using PivotTables
Percentiles and Quartiles
Cross‐Tabulations
Descriptive Statistical Measures
Populations and Samples
Statistical Notation
Measures of Location: Mean, Median, Mode, and Midrange
Using Measures of Location in Business Decisions
Measures of Dispersion: Range, Interquartile Range, Variance, and Standard Deviation
Chebyshev’s Theorem and the Empirical Rules
Standardized Values (Z‐Scores)
Coefficient of Variation
Measures of Shape
Excel Descriptive Statistics Tool
Computing Descriptive Statistics for Frequency Distributions
Descriptive Statistics for Categorical Data: The Proportion
Statistics in PivotTables
Measures of Association
Covariance
Correlation
Excel Correlation Tool
Outliers
Using Descriptive Statistics to Analyze Survey Data
Statistical Thinking in Business Decisions
Variability in Samples
Analytics in Practice: Applying Statistical Thinking to Detecting Financial Problems
Key Terms
Chapter 4 Technology Help
Problems and Exercises
Case: Drout Advertising Research Project
Case: Performance Lawn Equipment
Appendix A4: Additional Charts for Descriptive Statistics in Excel for Windows
Problems and Exercises
Chapter 5: Probability Distributions and Data Modeling
Learning Objectives
Basic Concepts of Probability
Experiments and Sample Spaces
Combinations and Permutations
Probability Definitions
Probability Rules and Formulas
Joint and Marginal Probability
Conditional Probability
Random Variables and Probability Distributions
Discrete Probability Distributions
Expected Value of a Discrete Random Variable
Using Expected Value in Making Decisions
Variance of a Discrete Random Variable
Bernoulli Distribution
Binomial Distribution
Poisson Distribution
Analytics in Practice: Using the Poisson Distribution for Modeling Bids on Priceline
Continuous Probability Distributions
Properties of Probability Density Functions
Uniform Distribution
Normal Distribution
The NORM.INV Function
Standard Normal Distribution
Using Standard Normal Distribution Tables
Exponential Distribution
Triangular Distribution
Data Modeling and Distribution Fitting
Goodness of Fit: Testing for Normality of an Empirical Distribution
Analytics in Practice: The value of Good Data Modeling in Advertising
Key Terms
Chapter 5 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Chapter 6: Sampling and Estimation
Learning Objectives
Statistical Sampling
Sampling Methods
Analytics in Practice: Using Sampling Techniques to Improve Distribution
Estimating Population Parameters
Unbiased Estimators
Errors in Point Estimation
Understanding Sampling Error
Sampling Distributions
Sampling Distribution of the Mean
Applying the Sampling Distribution of the Mean
Interval Estimates
Confidence Intervals
Confidence Interval for the Mean with Known Population Standard Deviation
The t‐Distribution
Confidence Interval for the Mean with Unknown Population Standard Deviation
Confidence Interval for a Proportion
Additional Types of Confidence Intervals
Using Confidence Intervals for Decision Making
Data Visualization for Confidence Interval Comparison
Prediction Intervals
Confidence Intervals and Sample Size
Key Terms
Chapter 6 Technology Help
Problems and Exercises
Case: Drout Advertising Research Project
Case: Performance Lawn Equipment
Chapter 7: Statistical Inference
Learning Objectives
Hypothesis Testing
Hypothesis‐Testing Procedure
One‐Sample Hypothesis Tests
Understanding Potential Errors in Hypothesis Testing
Selecting the Test Statistic
Finding Critical Values and Drawing a Conclusion
Two‐Tailed Test of Hypothesis for the Mean
Summary of One‐Sample Hypothesis Tests for the Mean
p‐Values
One‐Sample Tests for Proportions
Confidence Intervals and Hypothesis Tests
An Excel Template for One‐Sample Hypothesis Tests
Two‐Sample Hypothesis Tests
Two‐Sample Tests for Differences in Means
Two‐Sample Test for Means with Paired Samples
Two‐Sample Test for Equality of Variances
Analysis of Variance (ANOVA)
Assumptions of ANOVA
Chi‐Square Test for Independence
Cautions in Using the Chi‐Square Test
Chi‐Square Goodness of Fit Test
Analytics in Practice: Using Hypothesis Tests and Business Analytics in a Help Desk Service Improvement Project
Key Terms
Chapter 7 Technology Help
Problems and Exercises
Case: Drout Advertising Research Project
Case: Performance Lawn Equipment
Part 3: Predictive Analytics
Chapter 8: Trendlines and Regression Analysis
Learning Objectives
Modeling Relationships and Trends in Data
Analytics in Practice: Using Predictive Trendline Models at Procter & Gamble
Simple Linear Regression
Finding the Best‐Fitting Regression Line
Using Regression Models for Prediction
Least‐Squares Regression
Simple Linear Regression with Excel
Regression as Analysis of Variance
Testing Hypotheses for Regression Coefficients
Confidence Intervals for Regression Coefficients
Residual Analysis and Regression Assumptions
Checking Assumptions
Multiple Linear Regression
Analytics in Practice: Using Linear Regression and Interactive Risk Simulators to Predict Performance at Aramark
Building Good Regression Models
Correlation and Multicollinearity
Practical Issues in Trendline and Regression Modeling
Regression with Categorical Independent Variables
Categorical Variables with More Than Two Levels
Regression Models with Nonlinear Terms
Key Terms
Chapter 8 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Chapter 9: Forecasting Techniques
Learning Objectives
Analytics in Practice: Forecasting Call‐Center Demand at L.L. Bean
Qualitative and Judgmental Forecasting
Historical Analogy
The Delphi Method
Indicators and Indexes
Statistical Forecasting Models
Forecasting Models for Stationary Time Series
Moving Average Models
Error Metrics and Forecast Accuracy
Exponential Smoothing Models
Forecasting Models for Time Series with a Linear Trend
Double Exponential Smoothing
Regression‐Based Forecasting for Time Series with a Linear Trend
Forecasting Time Series with Seasonality
Regression‐Based Seasonal Forecasting Models
Holt‐Winters Models for Forecasting Time Series with Seasonality and No Trend
Holt‐Winters Models for Forecasting Time Series with Seasonality and Trend
Selecting Appropriate Time‐Series‐Based Forecasting Models
Regression Forecasting with Causal Variables
The Practice of Forecasting
Analytics in Practice: Forecasting at NBCUniversal
Key Terms
Chapter 9 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Chapter 10: Introduction to Data Mining
Learning Objectives
The Scope of Data Mining
Cluster Analysis
Measuring Distance Between Objects
Normalizing Distance Measures
Clustering Methods
Classification
An Intuitive Explanation of Classification
Measuring Classification Performance
Classification Techniques
Association
Cause‐and‐Effect Modeling
Analytics In Practice: Successful Business Applications of Data Mining
Key Terms
Chapter 10 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Chapter 11: Spreadsheet Modeling and Analysis
Learning Objectives
Analytics in Practice: Using Spreadsheet Modeling and Analysis at Nestlé
Model‐Building Strategies
Building Models Using Logic and Business Principles
Building Models Using Influence Diagrams
Building Models Using Historical Data
Model Assumptions, Complexity, and Realism
Implementing Models on Spreadsheets
Spreadsheet Design
Spreadsheet Quality
Data Validation
Analytics in Practice: Spreadsheet Engineering at Procter & Gamble
Descriptive Spreadsheet Models
Staffing Decisions
Single‐Period Purchase Decisions
Overbooking Decisions
Analytics in Practice: Using an Overbooking Model at a Student Health Clinic
Retail Markdown Decisions
Predictive Spreadsheet Models
New Product Development Model
Cash Budgeting
Retirement Planning
Project Management
Prescriptive Spreadsheet Models
Portfolio Allocation
Locating Central Facilities
Job Sequencing
Analyzing Uncertainty and Model Assumptions
What‐If Analysis
Data Tables
Scenario Manager
Goal Seek
Key Terms
Chapter 11 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Chapter 12: Simulation and Risk Analysis
Learning Objectives
Monte Carlo Simulation
Random Sampling from Probability Distributions
Generating Random Variates using Excel Functions
Discrete Probability Distributions
Uniform Distributions
Exponential Distributions
Normal Distributions
Binomial Distributions
Triangular Distributions
Monte Carlo Simulation in Excel
Profit Model Simulation
New Product Development
Retirement Planning
Single‐Period Purchase Decisions
Overbooking Decisions
Project Management
Analytics in Practice: Implementing Large‐Scale Monte Carlo Spreadsheet Models
Dynamic Systems Simulation
Simulating Waiting Lines
Analytics in Practice: Using Systems Simulation for Agricultural Product Development
Key Terms
Chapter 12 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Part 4: Prescriptive Analytics
Chapter 13: Linear Optimization
Learning Objectives
Optimization Models
Analytics in Practice: Using Optimization Models for Sales Planning at NBC
Developing Linear Optimization Models
Identifying Decision Variables, the Objective, and Constraints
Developing a Mathematical Model
More About Constraints
Implementing Linear Optimization Models on Spreadsheets
Excel Functions to Avoid in Linear Optimization
Solving Linear Optimization Models
Solver Answer Report
Graphical Interpretation of Linear Optimization with Two Variables
How Solver Works
How Solver Creates Names in Reports
Solver Outcomes and Solution Messages
Unique Optimal Solution
Alternative (Multiple) Optimal Solutions
Unbounded Solution
Infeasibility
Applications of Linear Optimization
Blending Models
Dealing with Infeasibility
Portfolio Investment Models
Scaling Issues in Using Solver
Transportation Models
Multiperiod Production Planning Models
Multiperiod Financial Planning Models
Analytics in Practice: Linear Optimization in Bank Financial Planning
Key Terms
Chapter 13 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Chapter 14: Integer and Nonlinear Optimization
Learning Objectives
Integer Linear Optimization Models
Models with General Integer Variables
Workforce‐Scheduling Models
Alternative Optimal Solutions
Models with Binary Variables
Using Binary Variables to Model Logical Constraints
Applications in Supply Chain Optimization
Analytics in Practice: Supply Chain Optimization at Procter & Gamble
Nonlinear Optimization Models
A Nonlinear Pricing Decision Model
Quadratic Optimization
Practical Issues Using Solver for Nonlinear Optimization
Analytics in Practice: Applying Nonlinear Optimization at Prudential Securities
Non‐Smooth Optimization
Evolutionary Solver
Evolutionary Solver for Sequencing and Scheduling Models
The Traveling Salesperson Problem
Key Terms
Chapter 14 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Chapter 15: Optimization Analytics
Learning Objectives
What‐If Analysis for Optimization Models
Solver Sensitivity Report
Using the Sensitivity Report
Degeneracy
Interpreting Solver Reports for Nonlinear Optimization Models
Models with Bounded Variables
Auxiliary Variables for Bound Constraints
What‐If Analysis for Integer Optimization Models
Visualization of Solver Reports
Using Sensitivity Information Correctly
Key Terms
Chapter 15 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Part 5: Making Decisions
Chapter 16: Decision Analysis
Learning Objectives
Formulating Decision Problems
Decision Strategies Without Outcome Probabilities
Decision Strategies for a Minimize Objective
Decision Strategies for a Maximize Objective
Decisions with Conflicting Objectives
Decision Strategies with Outcome Probabilities
Average Payoff Strategy
Expected Value Strategy
Evaluating Risk
Decision Trees
Decision Trees and Risk
Sensitivity Analysis in Decision Trees
The Value of Information
Decisions with Sample Information
Bayes’s Rule
Utility and Decision Making
Constructing a Utility Function
Exponential Utility Functions
Analytics in Practice: Using Decision Analysis in Drug Development
Key Terms
Chapter 16 Technology Help
Problems and Exercises
Case: Performance Lawn Equipment
Appendix A
Glossary
Index
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z
Business Analytics
This page intentionally left blank
Business Analytics Methods, Models, and Decisions James R. Evans University of Cincinnati THIRD EDITION GLOBAL EDITION
Harlow, England • London • New York • Boston • San Francisco • Toronto • Sydney • Dubai • Singapore • Hong Kong Tokyo • Seoul • Taipei • New Delhi • Cape Town • São Paulo • Mexico City • Madrid • Amsterdam • Munich • Paris • Milan
Please contact https://support.pearson.com/getsupport/s/contactsupport with any queries on this content Cover image by Allies Interactive / Shutterstock. Microsoft and/or its respective suppliers make no representations about the suitability of the information contained in the documents and related graphics published as part of the services for any purpose. All such documents and related graphics are provided “as is” without warranty of any kind. Microsoft and/or its respective suppliers hereby disclaim all warranties and conditions with regard to this information, including all warranties and conditions of merchantability, whether express, implied or statutory, fitness for a particular purpose, title and noninfringement. In no event shall Microsoft and/or its respective suppliers be liable for any special, indirect or consequential damages or any damages whatsoever resulting from loss of use, data or profits, whether in an action of contract, negligence or other tortious action, arising out of or in connection with the use or performance of information available from the services. The documents and related graphics contained herein could include technical inaccuracies or typographical errors. Changes are periodically added to the information herein. Microsoft and/or its respective suppliers may make improvements and/or changes in the product(s) and/or the program(s) described herein at any time. Partial screen shots may be viewed in full within the software version specified. Microsoft® and Windows® are registered trademarks of the Microsoft Corporation in the United States and other countries. This book is not sponsored or endorsed by or affiliated with the Microsoft Corporation. Pearson Education Limited KAO Two KAO Park Hockham Way Harlow Essex CM17 9SR United Kingdom and Associated Companies throughout the world Visit us on the World Wide Web at: www.pearsonglobaleditions.com © Pearson Education Limited 2021 The rights of James R. Evans, to be identified as the author of this work, has been asserted by him in accordance with the Copyright, Designs and Patents Act 1988. Authorized adaptation from the United States edition, entitled Business Analytics, 3rd Edition, ISBN 9780135231678 by James R. Evans, published by Pearson Education © 2020. PEARSON, ALWAYS LEARNING, and MYLAB are exclusive trademarks owned by Pearson Education, Inc. or its affiliates in the U.S. and/or other countries. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without either the prior written permission of the publisher or a license permitting restricted copying in the United Kingdom issued by the Copyright Licensing Agency Ltd, Saffron House, 6–10 Kirby Street, London EC1N 8TS. All trademarks used herein are the property of their respective owners. The use of any trademark in this text does not vest in the author or publisher any trademark ownership rights in such trademarks, nor does the use of such trademarks imply any affiliation with or endorsement of this book by such owners. For information regarding permissions, request forms, and the appropriate contacts within the Pearson Education Global Rights and Permissions department, please visit www.pearsoned.com/permissions/. This eBook is a standalone product and may or may not include all assets that were part of the print version. It also does not provide access to other Pearson digital products like MyLab and Mastering. The publisher reserves the right to remove any material in this eBook at any time. Print ISBN 10: 1292339063 Print ISBN 13: 9781292339061 eBook ISBN 13: 9781292339047 British Library CataloguinginPublication Data A catalogue record for this book is available from the British Library 1 20 Typeset in Times NR MT Pro by SPi Global
Brief Contents Preface 17 About the Author 25 Credits 27
Part 1 Foundations of Business Analytics Chapter 1 Chapter 2
Introduction to Business Analytics 29 Database Analytics 75
Part 2 Descriptive Analytics Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7
Data Visualization 113 Descriptive Statistics 143 Probability Distributions and Data Modeling 201 Sampling and Estimation 247 Statistical Inference 275
Part 3 Predictive Analytics Chapter 8 Chapter 9 Chapter 10 Chapter 11 Chapter 12
Trendlines and Regression Analysis 311 Forecasting Techniques 353 Introduction to Data Mining 383 Spreadsheet Modeling and Analysis 405 Simulation and Risk Analysis 451
Part 4 Prescriptive Analytics Chapter 13 Linear Optimization 493 Chapter 14 Integer and Nonlinear Optimization 551 Chapter 15 Optimization Analytics 593 Part 5 Making Decisions Chapter 16 Decision Analysis 631 Appendix A 661 Glossary 685 Index 693
5
This page intentionally left blank
Contents Preface 17 About the Author 25 Credits 27 Part 1: Foundations of Business Analytics
Chapter 1: Introduction to Business Analytics 29 Learning Objectives 29 What Is Business Analytics? 31
Using Business Analytics 32 • Impacts and Challenges 33
Evolution of Business Analytics 34
Analytic Foundations 34 • Modern Business Analytics 35 • Software Support and Spreadsheet Technology 37
Analytics in Practice: Social Media Analytics 38 Descriptive, Predictive, and Prescriptive Analytics 39 Analytics in Practice: Analytics in the Home Lending and Mortgage Industry 41 Data for Business Analytics 42 Big Data 44 • Data Reliability and Validity 44
Models in Business Analytics 45
Descriptive Models 47 • Predictive Models 49 • Prescriptive Models 50 • Model Assumptions 51 • Uncertainty and Risk 53
Problem Solving with Analytics 54
Recognizing a Problem 54 • Defining the Problem 54 • Structuring the Problem 55 • Analyzing the Problem 55 • Interpreting Results and Making a Decision 55 • Implementing the Solution 55
Analytics in Practice: Developing Effective Analytical Tools at HewlettPackard 56 Key Terms 57 • Chapter 1 Technology Help 57 • Problems and Exercises 57 • Case: Performance Lawn Equipment 59
Appendix A1: Basic Excel Skills 61 Excel Formulas and Addressing 62 Copying Formulas 63
Useful Excel Tips 63 Excel Functions 64
Basic Excel Functions 64 • Functions for Specific Applications 65 • Insert Function 66 • Date and Time Functions 67
Miscellaneous Excel Functions and Tools 68
Range Names 68 • VALUE Function 71 • Paste Special 71 • Concatenation 72 • Error Values 72
Problems and Exercises 73
7
8
Contents
Chapter 2: Database Analytics 75 Learning Objectives 75 Data Sets and Databases 77 Using Range Names in Databases 78
Analytics in Practice: Using Big Data to Monitor Water Usage in Cary, North Carolina 79 Data Queries: Tables, Sorting, and Filtering 79
Sorting Data in Excel 80 • Pareto Analysis 81 • Filtering Data 82 • Database Functions 84
Analytics in Practice: Discovering the Value of Database Analytics at Allders International 86 Logical Functions 87 Lookup Functions for Database Queries 89 Excel Template Design 92 Data Validation Tools 93 • Form Controls 95
PivotTables 98
PivotTable Customization 100 • Slicers 103
Key Terms 104 • Chapter 2 Technology Help 104 • Problems and Exercises 105 • Case: People’s Choice Bank 109 • Case: Drout Advertising Research Project 110
Part 2: Descriptive Analytics
Chapter 3: Data Visualization 113 Learning Objectives 113 The Value of Data Visualization 114 Tools and Software for Data Visualization 116
Analytics in Practice: Data Visualization for the New York City Police Department’s Domain Awareness System 116 Creating Charts in Microsoft Excel 116
Column and Bar Charts 117 • Data Label and Data Table Chart Options 118 • Line Charts 119 • Pie Charts 120 • Area Charts 121 • Scatter Charts and Orbit Charts 122 • Bubble Charts 123 • Combination Charts 124 • Radar Charts 125 • Stock Charts 125 • Charts from PivotTables 125 • Geographic Data 126
Other Excel Data Visualization Tools 126
Data Bars 126 • Color Scales 127 • Icon Sets 128 • Sparklines 129
Dashboards 131 Analytics in Practice: Driving Business Transformation with IBM Business Analytics 132 Key Terms 133 • Chapter 3 Technology Help 133 • Problems and Exercises 134 • Case: Performance Lawn Equipment 135
Appendix A3: Additional Tools for Data Visualization 136 Hierarchy Charts 136 Waterfall Charts 136 PivotCharts 138 Tableau 139 Problems and Exercises 141
Contents
9
Chapter 4: Descriptive Statistics 143 Learning Objectives 143 Analytics in Practice: Applications of Statistics in Health Care 145 Metrics and Data Classification 146 Frequency Distributions and Histograms 148
Frequency Distributions for Categorical Data 148 • Relative Frequency Distributions 149 • Frequency Distributions for Numerical Data 150 • Grouped Frequency Distributions 151 • Cumulative Relative Frequency Distributions 154 • Constructing Frequency Distributions Using PivotTables 155
Percentiles and Quartiles 157 CrossTabulations 158 Descriptive Statistical Measures 160
Populations and Samples 160 • Statistical Notation 161 • Measures of Location: Mean, Median, Mode, and Midrange 161 • Using Measures of Location in Business Decisions 163 • Measures of Dispersion: Range, Interquartile Range, Variance, and Standard Deviation 165 • Chebyshev’s Theorem and the Empirical Rules 168 • Standardized Values (ZScores) 170 • Coefficient of Variation 171 • Measures of Shape 172 • Excel Descriptive Statistics Tool 174
Computing Descriptive Statistics for Frequency Distributions 175 Descriptive Statistics for Categorical Data: The Proportion 177 Statistics in PivotTables 178 Measures of Association 179
Covariance 180 • Correlation 181 • Excel Correlation Tool 183
Outliers 184 Using Descriptive Statistics to Analyze Survey Data 186 Statistical Thinking in Business Decisions 187 Variability in Samples 188 Analytics in Practice: Applying Statistical Thinking to Detecting Financial Problems 190 Key Terms 191 • Chapter 4 Technology Help 192 • Problems and Exercises 193 • Case: Drout Advertising Research Project 198 • Case: Performance Lawn Equipment 198
Appendix A4: Additional Charts for Descriptive Statistics in Excel for Windows 199 Problems and Exercises 200
Chapter 5: Probability Distributions and Data Modeling 201 Learning Objectives 201 Basic Concepts of Probability 203
Experiments and Sample Spaces 203 • Combinations and Permutations 203 • Probability Definitions 205 • Probability Rules and Formulas 207 • Joint and Marginal Probability 208 • Conditional Probability 210
Random Variables and Probability Distributions 213 Discrete Probability Distributions 215
Expected Value of a Discrete Random Variable 216 • Using Expected Value in Making Decisions 217 • Variance of a Discrete Random Variable 219 • Bernoulli Distribution 219 • Binomial Distribution 220 • Poisson Distribution 221
Analytics in Practice: Using the Poisson Distribution for Modeling Bids on Priceline 223
10
Contents
Continuous Probability Distributions 224
Properties of Probability Density Functions 224 • Uniform Distribution 225 • Normal Distribution 227 • The NORM.INV Function 228 • Standard Normal Distribution 229 • Using Standard Normal Distribution Tables 230 • Exponential Distribution 231 • Triangular Distribution 232
Data Modeling and Distribution Fitting 233 Goodness of Fit: Testing for Normality of an Empirical Distribution 235
Analytics in Practice: The Value of Good Data Modeling in Advertising 237 Key Terms 238 • Chapter 5 Technology Help 238 • Problems and Exercises 239 • Case: Performance Lawn Equipment 245
Chapter 6: Sampling and Estimation 247 Learning Objectives 247 Statistical Sampling 248 Sampling Methods 249
Analytics in Practice: Using Sampling Techniques to Improve Distribution 251 Estimating Population Parameters 252
Unbiased Estimators 252 • Errors in Point Estimation 253 • Understanding Sampling Error 254
Sampling Distributions 256
Sampling Distribution of the Mean 256 • Applying the Sampling Distribution of the Mean 257
Interval Estimates 257
Confidence Intervals 258 • Confidence Interval for the Mean with Known Population Standard Deviation 259 • The tDistribution 260 • Confidence Interval for the Mean with Unknown Population Standard Deviation 261 • Confidence Interval for a Proportion 261 • Additional Types of Confidence Intervals 263
Using Confidence Intervals for Decision Making 263 Data Visualization for Confidence Interval Comparison 264
Prediction Intervals 265 Confidence Intervals and Sample Size 266 Key Terms 268 • Chapter 6 Technology Help 268 • Problems and Exercises 269 • Case: Drout Advertising Research Project 272 • Case: Performance Lawn Equipment 273
Chapter 7: Statistical Inference 275 Learning Objectives 275 Hypothesis Testing 276 HypothesisTesting Procedure 276
OneSample Hypothesis Tests 277
Understanding Potential Errors in Hypothesis Testing 278 • Selecting the Test Statistic 279 • Finding Critical Values and Drawing a Conclusion 280 • TwoTailed Test of Hypothesis for the Mean 282 • Summary of OneSample Hypothesis Tests for the Mean 283 • pValues 284 • OneSample Tests for Proportions 285 • Confidence Intervals and Hypothesis Tests 286 • An Excel Template for OneSample Hypothesis Tests 286
TwoSample Hypothesis Tests 287
Contents
11
TwoSample Tests for Differences in Means 288 • TwoSample Test for Means with Paired Samples 290 • TwoSample Test for Equality of Variances 292
Analysis of Variance (ANOVA) 294 Assumptions of ANOVA 296
ChiSquare Test for Independence 297
Cautions in Using the ChiSquare Test 299 • ChiSquare Goodness of Fit Test 300
Analytics in Practice: Using Hypothesis Tests and Business Analytics in a Help Desk Service Improvement Project 301 Key Terms 302 • Chapter 7 Technology Help 302 • Problems and Exercises 304 • Case: Drout Advertising Research Project 309 • Case: Performance Lawn Equipment 309
Part 3: Predictive Analytics
Chapter 8: Trendlines and Regression Analysis 311 Learning Objectives 311 Modeling Relationships and Trends in Data 313 Analytics in Practice: Using Predictive Trendline Models at Procter & Gamble 317 Simple Linear Regression 317 Finding the BestFitting Regression Line 319 • Using Regression Models for Prediction 319 • LeastSquares Regression 320 • Simple Linear Regression with Excel 322 • Regression as Analysis of Variance 324 • Testing Hypotheses for Regression Coefficients 325 • Confidence Intervals for Regression Coefficients 325
Residual Analysis and Regression Assumptions 326 Checking Assumptions 327
Multiple Linear Regression 329 Analytics in Practice: Using Linear Regression and Interactive Risk Simulators to Predict Performance at Aramark 332 Building Good Regression Models 334
Correlation and Multicollinearity 336 • Practical Issues in Trendline and Regression Modeling 338
Regression with Categorical Independent Variables 338 Categorical Variables with More Than Two Levels 341
Regression Models with Nonlinear Terms 343 Key Terms 345 • Chapter 8 Technology Help 345 • Problems and Exercises 346 • Case: Performance Lawn Equipment 350
Chapter 9: Forecasting Techniques 353 Learning Objectives 353 Analytics in Practice: Forecasting CallCenter Demand at L.L.Bean 354 Qualitative and Judgmental Forecasting 355
Historical Analogy 327 • The Delphi Method 355 • Indicators and Indexes 356
Statistical Forecasting Models 357 Forecasting Models for Stationary Time Series 359
Moving Average Models 359 • Error Metrics and Forecast Accuracy 361 • Exponential Smoothing Models 363
Forecasting Models for Time Series with a Linear Trend 366
Double Exponential Smoothing 366 • RegressionBased Forecasting for Time Series with a Linear Trend 368
12
Contents
Forecasting Time Series with Seasonality 369
RegressionBased Seasonal Forecasting Models 369 • HoltWinters Models for Forecasting Time Series with Seasonality and No Trend 371 • HoltWinters Models for Forecasting Time Series with Seasonality and Trend 373 • Selecting Appropriate TimeSeriesBased Forecasting Models 376
Regression Forecasting with Causal Variables 376 The Practice of Forecasting 377 Analytics in Practice: Forecasting at NBCUniversal 378 Key Terms 379 • Chapter 9 Technology Help 380 • Problems and Exercises 380 • Case: Performance Lawn Equipment 382
Chapter 10: Introduction to Data Mining 383 Learning Objectives 383 The Scope of Data Mining 384 Cluster Analysis 386
Measuring Distance Between Objects 387 • Normalizing Distance Measures 388 • Clustering Methods 388
Classification 390
An Intuitive Explanation of Classification 391 • Measuring Classification Performance 392 • Classification Techniques 393
Association 398 CauseandEffect Modeling 400
Analytics In Practice: Successful Business Applications of Data Mining 402 Key Terms 402 • Chapter 10 Technology Help 403 • Problems and Exercises 403 • Case: Performance Lawn Equipment 404
Chapter 11: Spreadsheet Modeling and Analysis 405 Learning Objectives 405 Analytics in Practice: Using Spreadsheet Modeling and Analysis at Nestlé 407 ModelBuilding Strategies 407
Building Models Using Logic and Business Principles 407 • Building Models Using Influence Diagrams 408 • Building Models Using Historical Data 409 • Model Assumptions, Complexity, and Realism 410
Implementing Models on Spreadsheets 410
Spreadsheet Design 411 • Spreadsheet Quality 412 • Data Validation 414
Analytics in Practice: Spreadsheet Engineering at Procter & Gamble 416 Descriptive Spreadsheet Models 416
Staffing Decisions 417 • SinglePeriod Purchase Decisions 418 • Overbooking Decisions 420
Analytics in Practice: Using an Overbooking Model at a Student Health Clinic 421 Retail Markdown Decisions 421
Predictive Spreadsheet Models 423
New Product Development Model 423 • Cash Budgeting 425 • Retirement Planning 426 • Project Management 426
Prescriptive Spreadsheet Models 429
Portfolio Allocation 429 • Locating Central Facilities 430 • Job Sequencing 432
Contents
13
Analyzing Uncertainty and Model Assumptions 434
WhatIf Analysis 434 • Data Tables 434 • Scenario Manager 437 • Goal Seek 438
Key Terms 440 • Chapter 11 Technology Help 441 • Problems and Exercises 442 • Case: Performance Lawn Equipment 449
Chapter 12: Simulation and Risk Analysis 451 Learning Objectives 451 Monte Carlo Simulation 453 Random Sampling from Probability Distributions 455 Generating Random Variates using Excel Functions 457
Discrete Probability Distributions 457 • Uniform Distributions 458 • Exponential Distributions 459 • Normal Distributions 459 • Binomial Distributions 461 • Triangular Distributions 461
Monte Carlo Simulation in Excel 463
Profit Model Simulation 463 • New Product Development 466 • Retirement Planning 468 • SinglePeriod Purchase Decisions 469 • Overbooking Decisions 472 • Project Management 472
Analytics in Practice: Implementing LargeScale Monte Carlo Spreadsheet Models 474 Dynamic Systems Simulation 475 Simulating Waiting Lines 477
Analytics in Practice: Using Systems Simulation for Agricultural Product Development 480 Key Terms 481 • Chapter 12 Technology Help 481 • Problems and Exercises 481 • Case: Performance Lawn Equipment 491
Part 4: Prescriptive Analytics
Chapter 13: Linear Optimization 493 Learning Objectives 493 Optimization Models 494 Analytics in Practice: Using Optimization Models for Sales Planning at NBC 496 Developing Linear Optimization Models 497
Identifying Decision Variables, the Objective, and Constraints 498 • Developing a Mathematical Model 499 • More About Constraints 500 • Implementing Linear Optimization Models on Spreadsheets 502 • Excel Functions to Avoid in Linear Optimization 503
Solving Linear Optimization Models 504
Solver Answer Report 506 • Graphical Interpretation of Linear Optimization with Two Variables 507
How Solver Works 513 How Solver Creates Names in Reports 514
Solver Outcomes and Solution Messages 515
Unique Optimal Solution 515 • Alternative (Multiple) Optimal Solutions 515 • Unbounded Solution 515 • Infeasibility 517
14
Contents
Applications of Linear Optimization 519
Blending Models 519 • Dealing with Infeasibility 520 • Portfolio Investment Models 521 • Scaling Issues in Using Solver 523 • Transportation Models 526 • Multiperiod Production Planning Models 529 • Multiperiod Financial Planning Models 533
Analytics in Practice: Linear Optimization in Bank Financial Planning 536 Key Terms 537 • Chapter 13 Technology Help 537 • Problems and Exercises 538 • Case: Performance Lawn Equipment 550
Chapter 14: Integer and Nonlinear Optimization 551 Learning Objectives 551 Integer Linear Optimization Models 552
Models with General Integer Variables 553 • WorkforceScheduling Models 556 • Alternative Optimal Solutions 559
Models with Binary Variables 561
Using Binary Variables to Model Logical Constraints 562 • Applications in Supply Chain Optimization 563
Analytics in Practice: Supply Chain Optimization at Procter & Gamble 567 Nonlinear Optimization Models 567
A Nonlinear Pricing Decision Model 567 • Quadratic Optimization 571 • Practical Issues Using Solver for Nonlinear Optimization 572
Analytics in Practice: Applying Nonlinear Optimization at Prudential Securities 573 NonSmooth Optimization 574 Evolutionary Solver 574 • Evolutionary Solver for Sequencing and Scheduling Models 577 • The Traveling Salesperson Problem 579
Key Terms 581 • Chapter 14 Technology Help 581 • Problems and Exercises 582 • Case: Performance Lawn Equipment 591
Chapter 15: Optimization Analytics 593 Learning Objectives 593 WhatIf Analysis for Optimization Models 594
Solver Sensitivity Report 595 • Using the Sensitivity Report 600 • Degeneracy 601 • Interpreting Solver Reports for Nonlinear Optimization Models 601
Models with Bounded Variables 603 Auxiliary Variables for Bound Constraints 606
WhatIf Analysis for Integer Optimization Models 609 Visualization of Solver Reports 611 Using Sensitivity Information Correctly 618 Key Terms 622 • Chapter 15 Technology Help 622 • Problems and Exercises 622 • Case: Performance Lawn Equipment 629
Part 5: Making Decisions
Chapter 16: Decision Analysis 631 Learning Objectives 631 Formulating Decision Problems 633 Decision Strategies Without Outcome Probabilities 634
Contents
15
Decision Strategies for a Minimize Objective 634 • Decision Strategies for a Maximize Objective 636 • Decisions with Conflicting Objectives 636
Decision Strategies with Outcome Probabilities 638
Average Payoff Strategy 638 • Expected Value Strategy 638 • Evaluating Risk 639
Decision Trees 640
Decision Trees and Risk 642 • Sensitivity Analysis in Decision Trees 645
The Value of Information 646 Decisions with Sample Information 647 • Bayes’s Rule 648 Utility and Decision Making 649
Constructing a Utility Function 650 • Exponential Utility Functions 653
Analytics in Practice: Using Decision Analysis in Drug Development 654 Key Terms 655 • Chapter 16 Technology Help 655 • Problems and Exercises 656 • Case: Performance Lawn Equipment 660
Online Supplements: Information about how to access and use Analytic Solver Basic are available for download at www.pearsonglobaleditions.com. Getting Started with Analytic Solver Using Advanced Regression Techniques in Analytic Solver Using Forecasting Techniques in Analytic Solver Using Data Mining in Analytic Solver Model Analysis in Analytic Solver Using Monte Carlo Simulation in Analytic Solver Using Linear Optimization in Analytic Solver Using Integer and Nonlinear Optimization in Analytic Solver Using Optimization Parameter Analysis in Analytic Solver Using Decision Trees in Analytic Solver Appendix A 661 Glossary 685 Index 693
This page intentionally left blank
Preface In 2007, Thomas H. Davenport and Jeanne G. Harris wrote a groundbreaking book, Competing on Analytics: The New Science of Winning (Boston: Harvard Business School Press). They described how many organizations are using analytics strategically to make better decisions and improve customer and shareholder value. Over the past several years, we have seen remarkable growth in analytics among all types of organizations. The Institute for Operations Research and the Management Sciences (INFORMS) noted that analytics software as a service is predicted to grow at three times the rate of other business segments in upcoming years.1 In addition, the MIT Sloan Management Review in collaboration with the IBM Institute for Business Value surveyed a global sample of nearly 3,000 executives, managers, and analysts.2 This study concluded that topperforming organizations use analytics five times more than lower performers, that improvement of information and analytics was a top priority in these organizations, and that many organizations felt they were under significant pressure to adopt advanced information and analytics approaches. Since these reports were published, the interest in and the use of analytics has grown dramatically. In reality, business analytics has been around for more than a halfcentury. Business schools have long taught many of the core topics in business analytics—statistics, data analysis, information and decision support systems, and management science. However, these topics have traditionally been presented in separate and independent courses and supported by textbooks with little topical integration. This book is uniquely designed to present the emerging discipline of business analytics in a unified fashion consistent with the contemporary definition of the field.
About the Book This book provides undergraduate business students and introductory graduate students with the fundamental concepts and tools needed to understand the role of modern business analytics in organizations, to apply basic business analytics tools in a spreadsheet environment, and to communicate with analytics professionals to effectively use and interpret analytic models and results for making better business decisions. We take a balanced, holistic approach in viewing business analytics from descriptive, predictive, and prescriptive perspectives that define the discipline.
1Anne Robinson, Jack Levis, and Gary Bennett, INFORMS News: INFORMS to Officially Join Analytics Movement. http://www.informs.org/ORMSToday/PublicArticles/OctoberVolume37Number5/ INFORMSNewsINFORMStoOfficiallyJoinAnalyticsMovement. 2“Analytics: The New Path to Value,” MIT Sloan Management Review Research Report, Fall 2010.
17
18
Preface
This book is organized in five parts. 1. Foundations of Business Analytics The first two chapters provide the basic foundations needed to understand business analytics and to manipulate data using Microsoft Excel. Chapter 1 provides an introduction to business analytics and its key concepts and terminology, and includes an appendix that reviews basic Excel skills. Chapter 2, Database Analytics, is a unique chapter that covers intermediate Excel skills, Excel template design, and PivotTables. 2. Descriptive Analytics Chapters 3 through 7 cover fundamental tools and methods of data analysis and statistics. These chapters focus on data visualization, descriptive statistical measures, probability distributions and data modeling, sampling and estimation, and statistical inference. We subscribe to the American Statistical Association’s recommendations for teaching introductory statistics, which include emphasizing statistical literacy and developing statistical thinking, stressing conceptual understanding rather than mere knowledge of procedures, and using technology for developing conceptual understanding and analyzing data. We believe these goals can be accomplished without introducing every conceivable technique into an 800–1,000 page book as many mainstream books currently do. In fact, we cover all essential content that the state of Ohio has mandated for undergraduate business statistics across all public colleges and universities. 3. Predictive Analytics In this section, Chapters 8 through 12 develop approaches for applying trendlines and regression analysis, forecasting, introductory data mining techniques, building and analyzing models on spreadsheets, and simulation and risk analysis. 4. Prescriptive Analytics Chapters 13 and 14 explore linear, integer, and nonlinear optimization models and applications. Chapter 15, Optimization Analytics, focuses on whatif and sensitivity analysis in optimization, and visualization of Solver reports. 5. Making Decisions Chapter 16 focuses on philosophies, tools, and techniques of decision analysis.
Changes to the Third Edition The third edition represents a comprehensive revision that includes many significant changes. The book now relies only on native Excel, and is independent of platforms, allowing it to be used easily by students with either PC or Mac computers. These changes provide students with enhanced Excel skills and basic understanding of fundamental concepts. Analytic Solver is no longer integrated directly in the book, but is illustrated in online supplements to facilitate revision as new software updates may occur. These supplements plus information regarding how to access Analytic Solver may be accessed at http://www.pearsonglobaleditions.com. Key changes to this edition are as follows: ■■
Also available for purchase (separately) is MyLab Statistics, a teaching and learning platform that empowers you to reach every student. By combining trusted author content with digital tools and a flexible platform, MyLab personalizes the
Preface
■■
■■ ■■
■■ ■■ ■■ ■■ ■■
■■ ■■ ■■
■■
19
learning experience and improves results for each student. For example, new Excel and StatCrunch Projects help students develop business decisionmaking skills. Each chapter now includes a short section called Technology Help, which provides useful summaries of key Excel functions and procedures, and the use of supplemental software including StatCrunch and Analytic Solver Basic. Chapter 1 includes an Appendix reviewing basic Excel skills, which will be used throughout the book. Chapter 2, Database Analytics, is a new chapter derived from the second edition that focuses on applications of Excel functions and techniques for dealing with spreadsheet data, including a new section on Excel template design. Chapter 3, Data Visualization, includes a new Appendix illustrating Excel tools for Windows and a brief overview of Tableau. Chapter 5, Probability Distributions and Data Modeling, includes a new section on Combinations and Permutations. Chapter 6, Sampling and Estimation, provides a discussion of using data visualization for confidence interval comparison. Chapter 9, Forecasting Techniques, now includes Excel approaches for double exponential smoothing and HoltWinters models for seasonality and trend. Chapter 10, Introduction to Data Mining, has been completely rewritten to illustrate simple data mining techniques that can be implemented on spreadsheets using Excel. Chapter 11, Spreadsheet Modeling and Analysis, is now organized along the analytic classification of descriptive, predictive, and prescriptive modeling. Chapter 12 has been rewritten to apply MonteCarlo simulation using only Excel, with an additional section of systems simulation concepts and approaches. Optimization topics have been reorganized into two chapters—Chapter 13, Linear Optimization, and Chapter 14, Integer and Nonlinear Optimization, which rely only on the Excelsupplied Solver. Chapter 15 is a new chapter called Optimization Analytics, which focuses on whatif and sensitivity analysis, and visualization of Solver reports; it also includes a discussion of how Solver handles models with bounded variables. In addition, we have carefully checked, and revised as necessary, the text and problems for additional clarity. We use major section headings in each chapter and tie these clearly to the problems and exercises, which have been revised and updated throughout the book. At the end of each section we added several “Check Your Understanding” questions that provide a basic review of fundamental concepts to improve student learning. Finally, new Analytics in Practice features have been incorporated into several chapters.
Features of the Book ■■ ■■
■■ ■■
Chapter Section Headings—with “Check Your Understanding” questions that provide a means to review fundamental concepts. Numbered Examples—numerous, short examples throughout all chapters illustrate concepts and techniques and help students learn to apply the techniques and understand the results. “Analytics in Practice”—at least one per chapter, this feature describes real applications in business. Learning Objectives—lists the goals the students should be able to achieve after studying the chapter.
20
Preface
■■
■■ ■■ ■■
Key Terms—bolded within the text and listed at the end of each chapter, these words will assist students as they review the chapter and study for exams. Key terms and their definitions are contained in the glossary at the end of the book. EndofChapter Problems and Exercises—clearly tied to sections in each chapter, these help to reinforce the material covered through the chapter. Integrated Cases—allow students to think independently and apply the relevant tools at a higher level of learning. Data Sets and Excel Models—used in examples and problems and are available to students at www.pearsonglobaleditions.com.
Software Support Technology Help sections in each chapter provide additional support to students for using Excel functions and tools, Tableau, and StatCrunch. Online supplements provide detailed information and examples for using Analytic Solver Basic, which provides more powerful tools for data mining, MonteCarlo simulation, optimization, and decision analysis. These can be used at the instructor’s discretion, but are not necessary to learn the fundamental concepts that are implemented using Excel. Instructions for obtaining licenses for Analytic Solver Basic can be found on the book’s website, http://www.pearsonglobaleditions.com.
To the Students To get the most out of this book, you need to do much more than simply read it! Many examples describe in detail how to use and apply various Excel tools or addins. We highly recommend that you work through these examples on your computer to replicate the outputs and results shown in the text. You should also compare mathematical formulas with spreadsheet formulas and work through basic numerical calculations by hand. Only in this fashion will you learn how to use the tools and techniques effectively, gain a better understanding of the underlying concepts of business analytics, and increase your proficiency in using Microsoft Excel, which will serve you well in your future career. Visit the companion Web site (www.pearsonglobaleditions.com) for access to the following: ■■
■■
Online Files: Data Sets and Excel Models—files for use with the numbered examples and the endofchapter problems. (For easy reference, the relevant file names are italicized and clearly stated when used in examples.) Online Supplements for Analytic Solver Basic: Online supplements describing the use of Analytic Solver that your instructor might use with selected chapters.
To the Instructors MyLab Statistics is now available with Evans “Business Analytics” 3e: MyLab™ Statistics is the teaching and learning platform that empowers instructors to reach every student. Teach your course your way with a flexible platform. Collect, crunch, and communicate with data in StatCrunch®, an integrated Webbased statistical software. Empower each learner with personalized and interactive practice. Tailor your course to your students’ needs with enhanced reporting features. Available with the complete eText, accessible anywhere with the Pearson eText app.
Preface
21
Instructor’s Resource Center—Reached through a link at www.pearsonglobaleditions. com, the Instructor’s Resource Center contains the electronic files for the complete Instructor’s Solutions Manual, PowerPoint lecture presentations, and the Test Item File. ■■
■■
■■
■■ ■■
Register, redeem, log in at www.pearsonglobaleditions.com: instructors can access a variety of print, media, and presentation resources that are available with this book in downloadable digital format. Instructor’s Solutions Manual—The Instructor’s Solutions Manual, updated and revised for the second edition by the author, includes Excelbased solutions for all endofchapter problems, exercises, and cases. PowerPoint presentations—The PowerPoint slides, revised and updated by the author, provide an instructor with individual lecture outlines to accompany the text. The slides include nearly all of the figures, tables, and examples from the text. Instructors can use these lecture notes as they are or can easily modify the notes to reflect specific presentation needs. Test Bank—The TestBank is prepared by Paolo Catasti from Virginia Commonwealth University. Need help? Pearson Education’s dedicated technical support team is ready to assist instructors with questions about the media supplements that accompany this text. The supplements are available to adopting instructors. Detailed descriptions are provided at the Instructor’s Resource Center.
Acknowledgments I would like to thank the staff at Pearson Education for their professionalism and dedication to making this book a reality. In particular, I want to thank Angela Montoya, Kathleen Manley, Karen Wernholm, Kaylee Carlson, Jean Choe, Bob Carroll, and Patrick Barbera. I would also like to thank Gowri Duraiswamy at SPI, and accuracy and solutions checker Jennifer Blue for their outstanding contributions to producing this book. I also want to acknowledge Daniel Fylstra and his staff at Frontline Systems for working closely with me on providing Analytic Solver Basic as a supplement with this book. If you have any suggestions or corrections, please contact the author via email at [email protected] James R. Evans Department of Operations, Business Analytics, and Information Systems University of Cincinnati Cincinnati, Ohio
Global Edition Acknowledgments Pearson would like to thank Alicia Tan Yiing Fei, Taylor’s University Malaysia; Dániel Kehl, University of Pécs; and Roland Baczur, University of Pécs for their contribution to the Global Edition.
Get the Most Out of
MyLab Statistics
Statistics courses are continuously evolving to help today’s students succeed. It’s more challenging than ever to support students with a wide range of backgrounds, learner styles, and math anxieties. The flexibility to build a course that fits instructors’ individual course formats—with a variety of content options and multimedia resources all in one place—has made MyLab Statistics the marketleading solution for teaching and learning mathematics since its inception.
78% of students say MyLab Statistics helped them learn their course content.*
Teach your course with a consistent author voice With marketleading author content options, your course can fit your style. Pearson offers the widest variety of content options, addressing a range of approaches and learning styles, authored by thought leaders across the business and math curriculum. MyLab™ Statistics is tightly integrated with each author’s style, offering a range of authorcreated multimedia resources, so your students have a consistent experience. Thanks to feedback from instructors and students from more than 10,000 institutions, MyLab Statistics continues to transform—delivering new content, innovative learning resources, and platform updates to support students and instructors, today and in the future.
*Source: 2018 Student Survey, n 31,721
pearson.com/mylab/statistics
Resources for Success
MyLab Statistics Online Course for Business Analytics by James R. Evans MyLab™ Statistics is available to accompany Pearson’s market leading text offerings. To give students a consistent tone, voice, and teaching method each text’s flavor and approach is tightly integrated throughout the accompanying MyLab Statistics course, making learning the material as seamless as possible.
Enjoy hands off grading with Excel Projects Using proven, fieldtested technology, autograded Excel Projects let instructors seamlessly integrate Microsoft Excel content into the course without manually grading spreadsheets. Students can practice important statistical skills in Excel, helping them master key concepts and gain proficiency with the program.
StatCrunch
StatCrunch, Pearson’s powerful webbased statistical software, instructors and students can access tens of thousands of data sets including those from the textbook, perform complex analyses, and generate compelling reports. StatCrunch is integrated directly into MyLab Statistics or available as a standalone product. To learn more, go to www.statcrunch.com on any laptop, tablet, or smartphone.
Technology Tutorials and Study Cards
MyLab makes learning and using a variety of statistical software programs as seamless and intuitive as possible. Download data sets from the text and MyLab exercises directly into Excel. Students can also access instructional support tools including tutorial videos, study cards, and manuals for a variety of statistical software programs including StatCrunch, Excel, Minitab, JMP, R, SPSS, and TI83/84 calculators.
pearson.com/mylab/statistics
This page intentionally left blank
About the Author James R. Evans Professor Emeritus, University of Cincinnati, Lindner College of Business James R. Evans is Professor Emeritus in the Department of Operations, Business Analytics, and Information Systems in the College of Business at the University of Cincinnati. He holds BSIE and MSIE degrees from Purdue and a PhD in Industrial and Systems Engineering from Georgia Tech. Dr. Evans has published numerous textbooks in a variety of business disciplines, including statistics, decision models, and analytics, simulation and risk analysis, network optimization, operations management, quality management, and creative thinking. He has published 100 papers in journals such as Management Science, IIE Transactions, Decision Sciences, Interfaces, the Journal of Operations Management, the Quality Management Journal, and many others, and wrote a series of columns in Interfaces on creativity in management science and operations research during the 1990s. He has also served on numerous journal editorial boards and is a pastpresident and Fellow of the Decision Sciences Institute. In 1996, he was an INFORMS Edelman Award Finalist as part of a project in supply chain optimization with Procter & Gamble that was credited with helping P&G save over $250,000,000 annually in their North American supply chain, and consulted on risk analysis modeling for Cincinnati 2012’s Olympic Games bid proposal. A recognized international expert on quality management, he served on the Board of Examiners and the Panel of Judges for the Malcolm Baldrige National Quality Award. Much of his research has focused on organizational performance excellence and measurement practices.
25
This page intentionally left blank
Credits Text Credits Chapter 3 page 129 Prem Thomas, MD and Seth Powsner, MD “Data Presentation for Quality Improvement”, 2005, AMIA. Appendix A page 661–663 National Institute of Standards and Technology.
Photo Credits Chapter 1 page 29 NAN728/Shutterstock page 56 hans12/Fotolia Chapter 2 page 75 NAN728/Shutterstock page 86 2jenn/Shutterstock Chapter 3 page 113 ESB Professional/Shutterstock Chapter 4 page 143 Nataliiap/Shutterstock page 191 langstrup/123RF Chapter 5 page 201 PeterVrabel/Shutterstock page 223 Fantasista/Fotolia page 237 Victor Correia/Shutterstock.com Chapter 6 page 247 Robert Brown Stock/Shutterstock page 252 Stephen Finn/ Shutterstock.com Chapter 7 page 275 Jirsak/Shutterstock page 301 Hurst Photo/Shutterstock Chapter 8 page 311 Luca Bertolli/123RF page 333 Gunnar Pippel/Shutterstock page 333 VectorIllustration/Shutterstock page 333 Claudio Divizia/Shutterstock page 333 Nataliia Natykach/Shutterstock Chapter 9 page 353 rawpixel/123RF page 379 Sean Pavone/Shutterstock Chapter 10 page 383 Laborant/Shutterstock page 402 Helder Almeida/Shutterstock Chapter 11 page 405 marekuliasz/Shutterstock page 416 Bryan Busovicki/Shutterstock page 421 Poprotskiy Alexey/Shutterstock Chapter 12 page 451 Stephen Rees/Shutterstock page 475 Vladitto/Shutterstock Chapter 13 page 493 Pinon Road/Shutterstock page 496 bizoo_n/Fotolia page 537 2jenn/Shutterstock Chapter 14 page 551 Jirsak/Shutterstock page 567 Kheng Guan Toh/Shutterstock Chapter 15 page 593 Alexander Orlov/Shutterstock Chapter 16 page 631 marekuliasz/Shutterstock page 655 SSokolov/Shutterstock Front Matter James R. Evans 27
This page intentionally left blank
CHAPTER
1
Introduction to Business Analytics
NAN728/Shutterstock
LEARNING OBJECTIVES ■■ ■■
■■
■■
■■
■■
After studying this chapter, you will be able to:
Define business analytics. Explain why analytics is important in today’s business environment. State some typical examples of business applications in which analytics would be beneficial. Summarize the evolution of business analytics and explain the concepts of business intelligence, operations research and management science, and decision support systems. Explain the difference between descriptive, predictive, and prescriptive analytics. State examples of how data are used in business.
■■
■■ ■■
■■ ■■
■■
■■
Explain the concept of a model and various ways a model can be characterized. Define and list the elements of a decision model. Illustrate examples of descriptive, predictive, and prescriptive models. Explain the difference between uncertainty and risk. Define the terms optimization, objective function, and optimal solution. Explain the difference between a deterministic and stochastic decision model. List and explain the steps in the problemsolving process.
30
Chapter 1 Introduction to Business Analytics
The purpose of this book is to provide you with a basic introduction to the concepts, methods, and models used in business analytics so that you will develop an appreciation not only for its capabilities to support and enhance business decisions, but also for the ability to use business analytics at an elementary level in your work. In this chapter, we introduce you to the field of business analytics and set the foundation for many of the concepts and techniques that you will learn. Let’s start with a rather innovative example. Most of you have likely been to a zoo, seen the animals, had something to eat, and bought some souvenirs. You probably wouldn’t think that managing a zoo is very difficult; after all, it’s just feeding and taking care of the animals, right? A zoo might be the last place that you would expect to find business analytics being used, but not anymore. The Cincinnati Zoo & Botanical Garden has been an “early adopter” and one of the first organizations of its kind to exploit business analytics.1 Despite generating more than twothirds of its budget through its own fundraising efforts, the zoo wanted to reduce its reliance on local tax subsidies even further by increasing visitor attendance and revenues from secondary sources such as membership, food, and retail outlets. The zoo’s senior management surmised that the best way to realize more value from each visit was to offer visitors a truly transformed customer experience. By using business analytics to gain greater insight into visitors’ behavior and tailoring operations to their preferences, the zoo expected to increase attendance, boost membership, and maximize sales. The project team—which consisted of consultants from IBM and Brightstar Partners, as well as senior executives from the zoo—began translating the organization’s goals into technical solutions. The zoo worked to create a business analytics platform that was capable of delivering the desired goals by combining data from ticketing and pointofsale systems throughout the zoo with membership information and geographical data gathered from the ZIP codes of all visitors. This enabled the creation of reports and dashboards that gave everyone from senior managers to zoo staff access to realtime information that helped them optimize operational management and transform the customer experience. By integrating weather forecast data, the zoo is now able to compare current forecasts with historic attendance and sales data, supporting better decision making for labor scheduling and inventory planning. Another area where the solution delivers new insight is food service. By opening food outlets at specific times of day when demand is highest (for example, keeping ice cream kiosks open in the 1
IBM Software Business Analtyics, “Cincinnati Zoo transforms customer experience and boosts profits,” © IBM Corporation 2012.
Chapter 1 Introduction to Business Analytics
31
final hour before the zoo closes), the zoo has been able to increase sales significantly. In addition, attendance and revenues have dramatically increased, resulting in annual return on investment of 411%. The business analytics initiative paid for itself within three months and delivers, on average, benefits of $738,212 per year. Specifically, ■■
The zoo has seen a 4.2% rise in ticket sales by targeting potential visitors who live in specific ZIP codes.
■■
Food revenues increased 25% by optimizing the mix of products on sale and adapting selling practices to match peak purchase times.
■■
Eliminating slowselling products and targeting visitors with specific promotions enabled an 18% increase in merchandise sales.
■■
The zoo was able to cut its marketing expenditure, saving $40,000 in the first year, and reduce advertising expenditure by 43% by eliminating ineffective campaigns and segmenting customers for more targeted marketing.
Because of the zoo’s success, other organizations such as Point Defiance Zoo & Aquarium in Tacoma, Washington, and History Colorado Center, a museum in Denver, have embarked on similar initiatives.
What Is Business Analytics? Everyone makes decisions. Individuals face personal decisions such as choosing a college or graduate program, making product purchases, selecting a mortgage instrument, and investing for retirement. Managers in business organizations make numerous decisions every day. Some of these decisions include what products to make and how to price them, where to locate facilities, how many people to hire, where to allocate advertising budgets, whether or not to outsource a business function or make a capital investment, and how to schedule production. Many of these decisions have significant economic consequences; moreover, they are difficult to make because of uncertain data and imperfect information about the future. Managers today no longer make decisions based on pure judgment and experience; they rely on factual data and the ability to manipulate and analyze data to supplement their intuition and experience, and to justify their decisions. What makes business decisions complicated today is the overwhelming amount of available data and information. Data to support business decisions—including those specifically collected by firms as well as through the Internet and social media such as Facebook—are growing exponentially and becoming increasingly difficult to understand and use. As a result, many companies have recently established analytics departments; for instance, IBM reorganized its consulting business and established a new 4,000person organization focusing on analytics. Companies are increasingly seeking business graduates with the ability to understand and use analytics. The demand for professionals with analytics expertise has skyrocketed, and many universities now have programs in analytics.2 2
Matthew J. Liberatore and Wenhong Luo, “The Analytics Movement: Implications for Operations Research,” Interfaces, 40, 4 (July–August 2010): 313–324.
32
Chapter 1 Introduction to Business Analytics
Business analytics, or simply analytics, is the use of data, information technology, statistical analysis, quantitative methods, and mathematical or computerbased models to help managers gain improved insight about their business operations and make better, factbased decisions. Business analytics is “a process of transforming data into actions through analysis and insights in the context of organizational decision making and problem solving.”3 Business analytics is supported by various tools such as Microsoft Excel and various Excel addins, commercial statistical software packages such as SAS or Minitab, and more complex business intelligence suites that integrate data with analytical software.
Using Business Analytics Tools and techniques of business analytics are used across many areas in a wide variety of organizations to improve the management of customer relationships, financial and marketing activities, human capital, supply chains, and many other areas. Leading banks use analytics to predict and prevent credit fraud. Investment firms use analytics to select the best client portfolios to manage risk and optimize return. Manufacturers use analytics for production planning, purchasing, and inventory management. Retailers use analytics to recommend products to customers and optimize marketing promotions. Pharmaceutical firms use analytics to get lifesaving drugs to market more quickly. The leisure and vacation industries use analytics to analyze historical sales data, understand customer behavior, improve Web site design, and optimize schedules and bookings. Airlines and hotels use analytics to dynamically set prices over time to maximize revenue. Even sports teams are using business analytics to determine both game strategy and optimal ticket prices.4 For example, teams use analytics to decide on ticket pricing, who to recruit and trade, what combinations of players work best, and what plays to run under different situations. Among the many organizations that use analytics to make strategic decisions and manage daytoday operations are Caesars Entertainment, the Cleveland Indians baseball, Phoenix Suns basketball, and New England Patriots football teams, Amazon.com, Procter & Gamble, United Parcel Service (UPS), and Capital One bank. It was reported that nearly all firms with revenues of more than $100 million are using some form of business analytics. Some common types of business decisions that can be enhanced by using analytics include ■■ ■■ ■■ ■■ ■■
3
pricing (for example, setting prices for consumer and industrial goods, government contracts, and maintenance contracts), customer segmentation (for example, identifying and targeting key customer groups in retail, insurance, and credit card industries), merchandising (for example, determining brands to buy, quantities, and allocations), location (for example, finding the best location for bank branches and ATMs, or where to service industrial equipment), supply chain design (for example, determining the best sourcing and transportation options and finding the best delivery routes),
Liberatore and Luo, “The Analytics Movement”. Jim Davis, “8 Essentials of Business Analytics,” in “Brain Trust—Enabling the Confident Enterprise with Business Analytics” (Cary, NC: SAS Institute, Inc., 2010): 27–29. www.sas.com/bareport 4
Chapter 1 Introduction to Business Analytics
■■ ■■
33
staffing (for example, ensuring the appropriate staffing levels and capabilities and hiring the right people—sometimes referred to as “people analytics”), health care (for example, scheduling operating rooms to improve utilization, improving patient flow and waiting times, purchasing supplies, predicting health risk factors),
and many others in operations management, finance, marketing, and human resources—in fact, in every discipline of business.5 Various research studies have discovered strong relationships between a company’s performance in terms of profitability, revenue, and shareholder return and its use of analytics. Topperforming organizations (those that outperform their competitors) are three times more likely to be sophisticated in their use of analytics than lower performers and are more likely to state that their use of analytics differentiates them from competitors.6 However, research has also suggested that organizations are overwhelmed by data and struggle to understand how to use data to achieve business results and that most organizations simply don’t understand how to use analytics to improve their businesses. Thus, understanding the capabilities and techniques of analytics is vital to managing in today’s business environment. So, no matter what your job position in an organization is or will be, the study of analytics will be quite important to your future success. You may find many uses in your everyday work for the Excelbased tools that we will study. You may not be skilled in all the technical nuances of analytics and supporting software, but you will, at the very least, be a consumer of analytics and work with analytics professionals to support your analyses and decisions. For example, you might find yourself on project teams with managers who know very little about analytics and analytics experts such as statisticians, programmers, and economists. Your role might be to ensure that analytics is used properly to solve important business problems.
Impacts and Challenges The benefits of applying business analytics can be significant. Companies report reduced costs, better risk management, faster decisions, better productivity, and enhanced bottomline performance such as profitability and customer satisfaction. For example, 1800Flowers.com used analytic software to target print and online promotions with greater accuracy; change prices and offerings on its Web site (sometimes hourly); and optimize its marketing, shipping, distribution, and manufacturing operations, resulting in a $50 million cost savings in one year.7 Business analytics is changing how managers make decisions.8 To thrive in today’s business world, organizations must continually innovate to differentiate themselves from competitors, seek ways to grow revenue and market share, reduce costs, retain existing customers and acquire new ones, and become faster and leaner. IBM suggests that traditional management 5
Thomas H. Davenport, “How Organizations Make Better Decisions,” edited excerpt of an article distributed by the International Institute for Analytics published in “Brain Trust—Enabling the Confident Enterprise with Business Analytics” (Cary, NC: SAS Institute, Inc., 2010): 8–11. www.sas.com/bareport 6 Thomas H. Davenport and Jeanne G. Harris, Competing on Analytics (Boston: Harvard Business School Press, 2007): 46; Michael S. Hopkins, Steve LaValle, Fred Balboni, Nina Kruschwitz, and Rebecca Shockley, “10 Data Points: Information and Analytics at Work,” MIT Sloan Management Review, 52, 1 (Fall 2010): 27–31. 7 Jim Goodnight, “The Impact of Business Analytics on Performance and Profitability,” in “Brain Trust— Enabling the Confident Enterprise with Business Analytics” (Cary, NC: SAS Institute, Inc., 2010): 4–7. www.sas.com/bareport 8 Analytics: The New Path to Value, a joint MIT Sloan Management Review and IBM Institute for Business Value study.
34
Chapter 1 Introduction to Business Analytics
approaches are evolving in today’s analyticsdriven environment to include more factbased decisions as opposed to judgment and intuition, more prediction rather than reactive decisions, and the use of analytics by everyone at the point where decisions are made rather than relying on skilled experts in a consulting group.9 Nevertheless, organizations face many challenges in developing analytics capabilities, including lack of understanding of how to use analytics, competing business priorities, insufficient analytical skills, difficulty in getting good data and sharing information, and not understanding the benefits versus perceived costs of analytics studies. Successful application of analytics requires more than just knowing the tools; it requires a highlevel understanding of how analytics supports an organization’s competitive strategy and effective execution that crosses multiple disciplines and managerial levels. In 2011, a survey by Bloomberg Businessweek Research Services and SAS concluded that business analytics was still in the “emerging stage” and was used only narrowly within business units, not across entire organizations. The study also noted that many organizations lacked analytical talent, and those that did have analytical talent often didn’t know how to apply the results properly. While analytics was used as part of the decisionmaking process in many organizations, most business decisions are still based on intuition.10 Today, business analytics has matured in many organizations, but many more opportunities still exist. These opportunities are reflected in the job market for analytics professionals, or “data scientists,” as some call them. McKinsey & Company suggested that there is a shortage of qualified data scientists.11
CHECK YOUR UNDERSTANDING 1. Explain why analytics is important in today’s business environment. 2. Define business analytics. 3. State three examples of how business analytics is used in organizations. 4. What are the key benefits of using business analytics? 5. What challenges do organizations face in using analytics?
Evolution of Business Analytics Analytical methods, in one form or another, have been used in business for more than a century. The core of business analytics consists of three disciplines: business intelligence and information systems, statistics, and modeling and optimization.
Analytic Foundations The modern evolution of analytics began with the introduction of computers in the late 1940s and their development through the 1960s and beyond. Early computers provided the ability to store and analyze data in ways that were either very difficult or impossible to do manually. This facilitated the collection, management, analysis, and reporting of data, which 9 “Business Analytics and Optimization for the Intelligent Enterprise” (April 2009). www.ibm.com/qbs/ intelligententerprise 10 Bloomberg Businessweek Research Services and SAS, “The Current State of Business Analytics: Where Do We Go From Here?” (2011). 11 Andrew Jennings, “What Makes a Good Data Scientist?” Analytics Magazine (July–August 2013): 8–13. www.analyticsmagazine.org
Chapter 1 Introduction to Business Analytics
35
is often called business intelligence (BI), a term that was coined in 1958 by an IBM researcher, Hans Peter Luhn.12 Business intelligence software can answer basic questions such as “How many units did we sell last month?” “What products did customers buy and how much did they spend?” “How many credit card transactions were completed yesterday?” Using BI, we can create simple rules to flag exceptions automatically; for example, a bank can easily identify transactions greater than $10,000 to report to the Internal Revenue Service.13 BI has evolved into the modern discipline we now call information systems (IS). Statistics has a long and rich history, yet only rather recently has it been recognized as an important element of business, driven to a large extent by the massive growth of data in today’s world. Google’s chief economist noted that statisticians surely have one of the best jobs.14 Statistical methods allow us to gain a richer understanding of data that goes beyond business intelligence reporting by not only summarizing data succinctly but also finding unknown and interesting relationships among the data. Statistical methods include the basic tools of description, exploration, estimation, and inference, as well as more advanced techniques like regression, forecasting, and data mining. Much of modern business analytics stems from the analysis and solution of complex decision problems using mathematical or computerbased models—a discipline known as operations research, or management science. Operations research (OR) was born from efforts to improve military operations prior to and during World War II. After the war, scientists recognized that the mathematical tools and techniques developed for military applications could be applied successfully to problems in business and industry. A significant amount of research was carried on in public and private think tanks during the late 1940s and through the 1950s. As the focus on business applications expanded, the term management science (MS) became more prevalent. Many people use the terms operations research and management science interchangeably, so the field became known as Operations Research/Management Science (OR/MS). Many OR/MS applications use modeling and optimization—techniques for translating real problems into mathematics, spreadsheets, or various computer languages, and using them to find the best (“optimal”) solutions and decisions. INFORMS, the Institute for Operations Research and the Management Sciences, is the leading professional society devoted to OR/MS and analytics and publishes a bimonthly magazine called Analytics (http://analyticsmagazine.org/). Digital subscriptions may be obtained free of charge at the Web site.
Modern Business Analytics Modern business analytics can be viewed as an integration of BI/IS, statistics, and modeling and optimization, as illustrated in Figure 1.1. While these core topics are traditional and have been used for decades, the uniqueness lies in their intersections. For example, data mining is focused on better understanding characteristics and patterns among variables in large databases using a variety of statistical and analytical tools. Many standard statistical tools as well as more advanced ones are used extensively in data mining. Simulation and risk analysis relies on spreadsheet models and statistical analysis to examine the impacts of uncertainty in estimates and their potential interaction with one another on the output variable of interest. 12
H. P. Luhn, “A Business Intelligence System.” IBM Journal (October 1958). Jim Davis, “Business Analytics: Helping You Put an Informed Foot Forward,” in “Brain Trust—Enabling the Confident Enterprise with Business Analytics,” (Cary, NC: SAS Institute, Inc., 2010): 4–7. www.sas .com/bareport 14 James J. Swain, “Statistical Software in the Age of the Geek,” Analytics Magazine (March April 2013): 48–55. 13
36
Chapter 1 Introduction to Business Analytics
▶▶ Figure
1.1
A Visual Perspective of Business Analytics
Data Mining
Statistics
Visualization Simulation and Risk
Business Intelligence/ Information Systems Decision Support Systems
Modeling and Optimization
Decision support systems (DSSs) began to evolve in the 1960s by combining business intelligence concepts with OR/MS models to create analyticalbased computer systems to support decision making. DSSs include three components: 1. Data management. The data management component includes databases for storing data and allows the user to input, retrieve, update, and manipulate data. 2. Model management. The model management component consists of various statistical tools and management science models and allows the user to easily build, manipulate, analyze, and solve models. 3. Communication system. The communication system component provides the interface necessary for the user to interact with the data and model management components.15 DSSs have been used for many applications, including pension fund management, portfolio management, workshift scheduling, global manufacturing and facility location, advertisingbudget allocation, media planning, distribution planning, airline operations planning, inventory control, library management, classroom assignment, nurse scheduling, blood distribution, water pollution control, skiarea design, policebeat design, and energy planning.16 A key feature of a DSS is the ability to perform whatif analysis—how specific combinations of inputs that reflect key assumptions will affect model outputs. Whatif analysis is also used to assess the sensitivity of optimization models to changes in data inputs and provide better insight for making good decisions. Perhaps the most useful component of business analytics, which makes it truly unique, is the center of Figure 1.1—visualization. Visualizing data and results of analyses provides a way of easily communicating data at all levels of a business and can reveal surprising patterns and relationships. Software such as IBM’s Cognos system exploits data visualization 15 William E. Leigh and Michael E. Doherty, Decision Support and Expert Systems (Cincinnati, OH: SouthWestern Publishing Co., 1986). 16 H. B. Eom and S. M. Lee, “A Survey of Decision Support System Applications (1971–April 1988),” Interfaces, 20, 3 (May–June 1990): 65–79.
Chapter 1 Introduction to Business Analytics
37
for query and reporting, data analysis, dashboard presentations, and scorecards linking strategy to operations. The Cincinnati Zoo, for example, has used this on an iPad to display hourly, daily, and monthly reports of attendance, food and retail location revenues and sales, and other metrics for prediction and marketing strategies. UPS uses telematics to capture vehicle data and display them to help make decisions to improve efficiency and performance. You may have seen a tag cloud (see the graphic at the beginning of this chapter), which is a visualization of text that shows words that appear more frequently with larger fonts.
Software Support and Spreadsheet Technology Many companies, such as IBM, SAS, and Tableau Software, have developed a variety of software and hardware solutions to support business analytics. For example, IBM’s Cognos Express, an integrated business intelligence and planning solution designed to meet the needs of midsize companies, provides reporting, analysis, dashboard, scorecard, planning, budgeting, and forecasting capabilities. It is made up of several modules, including Cognos Express Reporter, for selfservice reporting and ad hoc query; Cognos Express Advisor, for analysis and visualization; and Cognos Express Xcelerator, for Excelbased planning and business analysis. Information is presented to users in a context that makes it easy to understand; with an easytouse interface, users can quickly gain the insight they need from their data to make the right decisions and then take action for effective and efficient business optimization and outcome. SAS provides a variety of software that integrate data management, business intelligence, and analytics tools. SAS Analytics covers a wide range of capabilities, including predictive modeling and data mining, visualization, forecasting, optimization and model management, statistical analysis, text analytics, and more. Tableau Software provides simple drag and drop tools for visualizing data from spreadsheets and other databases. We encourage you to explore many of these products as you learn the basic principles of business analytics in this book. Although commercial software often have powerful features and capabilities, they can be expensive, generally require advanced training to understand and apply, and may work only on specific computer platforms. Spreadsheet software, on the other hand, is widely used across all areas of business and used by nearly everyone. Spreadsheets are an effective platform for manipulating data and developing and solving models; they support powerful commercial addins and facilitate communication of results. Spreadsheets provide a flexible modeling environment and are particularly useful when the end user is not the designer of the model. Teams can easily use spreadsheets and understand the logic upon which they are built. Information in spreadsheets can easily be copied from spreadsheets into other documents and presentations. A recent survey identified more than 180 commercial spreadsheet products that support analytics efforts, including data management and reporting, data and modeldriven analytical techniques, and implementation. 17 Many organizations have used spreadsheets extremely effectively to support decision making in marketing, finance, and operations. Some illustrative applications include the following:18 ■■ ■■
17
Analyzing supply chains (HewlettPackard) Determining optimal inventory levels to meet customer service objectives (Procter & Gamble)
Thomas A. Grossman, “Resources for Spreadsheet Analysts,” Analytics Magazine (May/June 2010): 8. www.analyticsmagazine.org 18 Larry J. LeBlanc and Thomas A. Grossman, “Introduction: The Use of Spreadsheet Software in the Application of Management Science and Operations Research,” Interfaces, 38, 4 (July–August 2008): 225–227.
38
Chapter 1 Introduction to Business Analytics
■■ ■■ ■■ ■■ ■■
Selecting internal projects (Lockheed Martin Space Systems) Planning for emergency clinics in response to a sudden epidemic or bioterrorism attack (Centers for Disease Control) Analyzing the default risk of a portfolio of real estate loans (Hypo International) Assigning medical residents to oncall and emergency rotations (University of Vermont College of Medicine) Performance measurement and evaluation (American Red Cross)
Some optional software packages for statistical applications that your instructor might use are SAS, Minitab, XLSTAT and StatCrunch. These provide many powerful procedures as alternatives or supplements to Excel. Spreadsheet technology has been influential in promoting the use and acceptance of business analytics. Spreadsheets provide a convenient way to manage data, calculations, and visual graphics simultaneously, using intuitive representations instead of abstract mathematical notation. Although the early applications of spreadsheets were primarily in accounting and finance, spreadsheets have developed into powerful generalpurpose managerial tools for applying techniques of business analytics. The power of analytics in a personal computing environment was noted decades ago by business consultants Michael Hammer and James Champy, who said, “When accessible data is combined with easytouse analysis and modeling tools, frontline workers—when properly trained—suddenly have sophisticated decisionmaking capabilities.”19
ANALYTICS IN PRACTICE: Social Media Analytics One of the emerging applications of analytics is helping businesses learn from social media and exploit social media data for strategic advantage.20 Using analytics, firms can integrate social media data with traditional data sources such as customer surveys, focus groups, and sales data; understand trends and customer perceptions of their products; and create informative reports to assist marketing managers and product designers. Social media analytics is useful in decision making in many business domains to understand how usergenerated content spreads and influences user interactions, how information is transmitted, and how it influences decisions. A review of research published in social media analytics provides numerous examples:21 ■■
The analysis of public responses from social media before, during, and after disasters, such as the 2010 Haiti earthquake and Hurricane Sandy in New York City
■■
■■
in 2012, has the potential to improve situational knowledge in emergency and disaster management practices. Social media platforms enable citizens’ engagement with politicians, governments, and other citizens. Studies have examined how voters discuss the candidates during an election, how candidates are adopting Twitter for campaigning and influencing conversations in the public space, and how presidential candidates in the United States used Twitter to engage people and identify the topics mentioned by candidates during their campaigns. Others have used analytics to track political preference by monitoring online popularity. In the entertainment industry, one study analyzed viewer ratings to predict the impact on revenue for upcoming movies. Another developed a web intelligence application to aggregate the news about popular TV serials and identify emerging storylines.
19 Michael Hammer and James Champy, Reengineering the Corporation (New York: HarperBusiness, 1993): 96. 20 Jim Davis, “Convergence—Taking Social Media from Talk to Action,” SASCOM (First Quarter 2011): 17. 21 Ashish K. Rathore, Arpan K. Kar, and P. Vigneswara Ilavarasana, “Social Media Analytics: Literature Review and Directions for Future Research,” Decision Analysis, 14, 4 (December 2017): 229–249.
Chapter 1 Introduction to Business Analytics
■■
■■
Retail organizations monitor and analyze social media data about their own products and services and also about their competitors’ products and services to stay competitive. For instance, one study analyzed different product features based on rankings from users’ online reviews. The integration of social media application and health care leads to better patient management
39
and empowerment. One researcher classified various online health communities, such as a diabetes patients’ community, using posts from WebMD.com. Another analyzed physical activity–related tweets for a better understanding of physical activity behaviors. To predict the spread of influenza, one researcher developed a forecasting approach using flurelated tweets.
In this book, we use Microsoft Excel as the primary platform for implementing analytics. In the Chapter 1 Appendix, we review some key Excel skills that you should have before moving forward in this book. The main chapters in this book are designed using Excel 2016 for Windows or Excel 2016 for Mac. Earlier versions of Excel do not have all the capabilities that we use in this book. In addition, some key differences exist between Windows and Mac versions that we will occasionally point out. Thus, some Excel tools that we will describe in chapter appendixes require you to use Excel for Windows, Office 365, or Google Sheets, and will not run on Excel for Mac; these are optional to learn, and are not required for any examples or problems. Your instructor may use optional software, such as XLSTAT and StatCrunch, which are provided by the publisher (Pearson), or Analytic Solver, which is described in online supplements to this book. CHECK YOUR UNDERSTANDING 1. Provide two examples of questions that business intelligence can address. 2. How do statistical methods enhance business intelligence reporting? 3. What is operations research/management science? 4. How does modern business analytics integrate traditional disciplines of BI, statistics, and modeling/optimization? 5. What are the components of a decision support system?
Descriptive, Predictive, and Prescriptive Analytics Business analytics begins with the collection, organization, and manipulation of data and is supported by three major components:22 1. Descriptive analytics. Most businesses start with descriptive analytics—the use of data to understand past and current business performance and make informed decisions. Descriptive analytics is the most commonly used and most wellunderstood type of analytics. These techniques categorize, characterize, consolidate, and classify data to convert them into useful information for the purposes of understanding and analyzing business performance. Descriptive 22
Parts of this section are adapted from Irv Lustig, Brenda Dietric, Christer Johnson, and Christopher Dziekan, “The Analytics Journey,” Analytics (November/December 2010). http://analyticsmagazine.org/ novemberdecember2010tableofcontents/
40
Chapter 1 Introduction to Business Analytics
analytics summarizes data into meaningful charts and reports, for example, about budgets, sales, revenues, or cost. This process allows managers to obtain standard and customized reports and then drill down into the data and make queries to understand the impact of an advertising campaign, such as reviewing business performance to find problems or areas of opportunity, and identifying patterns and trends in data. Typical questions that descriptive analytics helps answer are “How much did we sell in each region?” “What was our revenue and profit last quarter?” “How many and what types of complaints did we resolve?” “Which factory has the lowest productivity?” Descriptive analytics also helps companies to classify customers into different segments, which enables them to develop specific marketing campaigns and advertising strategies. 2. Predictive analytics. Predictive analytics seeks to predict the future by examining historical data, detecting patterns or relationships in these data, and then extrapolating these relationships forward in time. For example, a marketer might wish to predict the response of different customer segments to an advertising campaign, a commodities trader might wish to predict shortterm movements in commodities prices, or a skiwear manufacturer might want to predict next season’s demand for skiwear of a specific color and size. Predictive analytics can predict risk and find relationships in data not readily apparent with traditional analyses. Using advanced techniques, predictive analytics can help detect hidden patterns in large quantities of data, and segment and group data into coherent sets to predict behavior and detect trends. For instance, a bank manager might want to identify the most profitable customers, predict the chances that a loan applicant will default, or alert a credit card customer to a potential fraudulent charge. Predictive analytics helps to answer questions such as “What will happen if demand falls by 10% or if supplier prices go up 5%?” “What do we expect to pay for fuel over the next several months?” “What is the risk of losing money in a new business venture?” 3. Prescriptive analytics. Many problems, such as aircraft or employee scheduling and supply chain design, simply involve too many choices or alternatives for a human decision maker to effectively consider. Prescriptive analytics uses optimization to identify the best alternatives to minimize or maximize some objective. Prescriptive analytics is used in many areas of business, including operations, marketing, and finance. For example, we may determine the best pricing and advertising strategy to maximize revenue, the optimal amount of cash to store in ATMs, or the best mix of investments in a retirement portfolio to manage risk. Prescriptive analytics addresses questions such as “How much should we produce to maximize profit?” “What is the best way of shipping goods from our factories to minimize costs?” “Should we change our plans if a natural disaster closes a supplier’s factory, and if so, by how much?” The mathematical and statistical techniques of predictive analytics can also be combined with prescriptive analytics to make decisions that take into account the uncertainty in the data. A wide variety of tools are used to support business analytics. These include ■■ ■■ ■■ ■■ ■■
Database queries and analysis “Dashboards” to report key performance measures Data visualization Statistical methods Spreadsheets and predictive models
Chapter 1 Introduction to Business Analytics
ANALYTICS IN PRACTICE: Analytics in the Home Lending and
Mortgage Industry23
Sometime during their lives, most Americans will receive a mortgage loan for a house or condominium. The process starts with an application. The application contains all pertinent information about the borrower that the lender will need. The bank or mortgage company then initiates a process that leads to a loan decision. It is here that key information about the borrower is provided by thirdparty providers. This information includes a credit report, verification of income, verification of assets, verification of employment, and an appraisal of the property. The result of the processing function is a complete loan file that contains all the information and documents needed to underwrite the loan, which is the next step in the process. Underwriting is where the loan application is evaluated for its risk. Underwriters evaluate whether the borrower can make payments on time, can afford to pay back the loan, and has sufficient collateral in the property to back up the loan. In the event the borrower defaults on their loan, the lender can sell the property to recover the amount of the loan. But if the amount of the loan is greater than the value of the property, then the lender cannot recoup their money. If the underwriting process indicates that the borrower is creditworthy and has the capacity to repay the loan and the value of the property in question is greater than the loan amount, then the loan is approved and will move to closing. Closing is the step where the borrower signs all the appropriate papers, agreeing to the terms of the loan. In reality, lenders have a lot of other work to do. First, they must perform a quality control review on a sample of the loan files that involves a manual examination of all the documents and information gathered. This process is designed to identify any mistakes that may have been made or information that is missing from the loan file. Because lenders do not have unlimited money to lend to borrowers, they frequently sell the loan to a third party so that they have fresh capital to lend to others. This occurs in what is called the secondary market. Freddie Mac and Fannie Mae are the two largest purchasers of mortgages in the secondary market. The final step in the process is servicing. Servicing includes all the activities associated with providing the customer service on the loan, like processing payments, managing property taxes held in escrow, and answering questions about the loan.
23
Contributed by Craig Zielazny, BlueNote Analytics, LLC.
In addition, the institution collects various operational data on the process to track its performance and efficiency, including the number of applications, loan types and amounts, cycle times (time to close the loan), bottlenecks in the process, and so on. Many different types of analytics are used: Descriptive analytics—This focuses on historical reporting, addressing such questions as ■■
■■ ■■
How many loan applications were taken in each of the past 12 months? What was the total cycle time from application to close? What was the distribution of loan profitability by credit score and loantovalue (LTV), which is the mortgage amount divided by the appraised value of the property?
Predictive analytics—Predictive modeling uses mathematical, spreadsheet, and statistical models and addresses questions such as ■■
■■
■■
What impact on loan volume will a given marketing program have? How many processors or underwriters are needed for a given loan volume? Will a given process change reduce cycle time?
Prescriptive analytics—This involves the use of simulation or optimization to drive decisions. Typical questions include ■■
■■
What is the optimal staffing to achieve a given profitability constrained by a fixed cycle time? What is the optimal product mix to maximize profit constrained by fixed staffing?
The mortgage market has become much more dynamic in recent years due to rising home values, falling interest rates, new loan products, and an increased desire by home owners to utilize the equity in their homes as a financial resource. This has increased the complexity and variability of the mortgage process and created an opportunity for lenders to proactively use the data that are available to them as a tool for managing their business. To ensure that the process is efficient, effective, and performed with quality, data and analytics are used every day to track what is done, who is doing it, and how long it takes.
41
42
Chapter 1 Introduction to Business Analytics
■■ ■■ ■■ ■■ ■■ ■■
Scenario and “whatif” analyses Simulation Forecasting Data and text mining Optimization Social media, Web, and text analytics
Although the tools used in descriptive, predictive, and prescriptive analytics are different, many applications involve all three. Here is a typical example in retail operations. EXAMPLE 1.1
Retail Markdown Decisions24
As you probably know from your shopping experiences, most department stores and fashion retailers clear their seasonal inventory by reducing prices. The key question they face is what prices should they set—and when should they set them—to meet inventory goals and maximize revenue? For example, suppose that a store has 100 bathing suits of a certain style that go on sale on April 1 and wants to sell all of them by the end of June. Over each week of the 12week selling season, they can make a decision to discount the price. They face two decisions: When to reduce the price, and by how much. This results in 24 decisions to
make. For a major national chain that may carry thousands of products, this can easily result in millions of decisions that store managers have to make. Descriptive analytics can be used to examine historical data for similar products, such as the number of units sold, price at each point of sale, starting and ending inventories, and special promotions, newspaper ads, direct marketing ads, and so on, to understand what the results of past decisions achieved. Predictive analytics can be used to predict sales based on pricing decisions. Finally, prescriptive analytics can be applied to find the best set of pricing decisions to maximize the total revenue.
CHECK YOUR UNDERSTANDING 1. Define descriptive analytics and provide two examples. 2. Define predictive analytics and provide two examples. 3. Define prescriptive analytics and provide two examples.
Data for Business Analytics Since the dawn of the electronic age and the Internet, both individuals and organizations have had access to an enormous wealth of data and information. Most data are collected through some type of measurement process, and consist of numbers (e.g., sales revenues) or textual data (e.g., customer demographics such as gender). Other data might be extracted from social media, online reviews, and even audio and video files. Information comes from analyzing data—that is, extracting meaning from data to support evaluation and decision making. Data are used in virtually every major function in a business. Modern organizations— which include not only forprofit businesses but also nonprofit organizations—need good data to support a variety of company purposes, such as planning, reviewing company performance, improving operations, and comparing company performance with competitors’ 24
Inspired by a presentation by Radhika Kulkarni, SAS Institute, “DataDriven Decisions: Role of perations Research in Business Analytics,” INFORMS Conference on Business Analytics and Operations O Research, April 10–12, 2011.
Chapter 1 Introduction to Business Analytics
43
or bestpractice benchmarks. Some examples of how data are used in business include the following: ■■
■■
■■
■■
■■
■■
■■
Annual reports summarize data about companies’ profitability and market share both in numerical form and in charts and graphs to communicate with shareholders. Accountants conduct audits to determine whether figures reported on a firm’s balance sheet fairly represent the actual data by examining samples (that is, subsets) of accounting data, such as accounts receivable. Financial analysts collect and analyze a variety of data to understand the contribution that a business provides to its shareholders. These typically include profitability, revenue growth, return on investment, asset utilization, operating margins, earnings per share, economic value added (EVA), shareholder value, and other relevant measures. Economists use data to help companies understand and predict population trends, interest rates, industry performance, consumer spending, and international trade. Such data are often obtained from external sources such as Standard & Poor’s Compustat data sets, industry trade associations, or government databases. Marketing researchers collect and analyze extensive customer data. These data often consist of demographics, preferences and opinions, transaction and payment history, shopping behavior, and much more. Such data may be collected by surveys, personal interviews, or focus groups, or from shopper loyalty cards. Operations managers use data on production performance, manufacturing quality, delivery times, order accuracy, supplier performance, productivity, costs, and environmental compliance to manage their operations. Human resource managers measure employee satisfaction, training costs, turnover, market innovation, training effectiveness, and skills development.
Data may be gathered from primary sources such as internal company records and business transactions, automated datacapturing equipment, and customer market surveys and from secondary sources such as government and commercial data sources, custom research providers, and online research. Perhaps the most important source of data today is data obtained from the Web. With today’s technology, marketers collect extensive information about Web behaviors, such as the number of page views, visitor’s country, time of view, length of time, origin and destination paths, products they searched for and viewed, products purchased, and what reviews they read. Using analytics, marketers can learn what content is being viewed most often, what ads were clicked on, who the most frequent visitors are, and what types of visitors browse but don’t buy. Not only can marketers understand what customers have done, but they can better predict what they intend to do in the future. For example, if a bank knows that a customer has browsed for mortgage rates and homeowner’s insurance, they can target the customer with homeowner loans rather than credit cards or automobile loans. Traditional Web data are now being enhanced with social media data from Facebook, cell phones, and even Internetconnected gaming devices. As one example, a home furnishings retailer wanted to increase the rate of sales for customers who browsed their Web site. They developed a large data set that covered more than 7,000 demographic, Web, catalog, and retail behavioral attributes for each customer. They used predictive analytics to determine how well a customer would respond to different email marketing offers and customized promotions to individual customers. This not only helped them to determine where to most effectively spend marketing resources but
44
Chapter 1 Introduction to Business Analytics
also doubled the response rate compared to previous marketing campaigns, with a p rojected and multimillion dollar increase in sales.25
Big Data Today, nearly all data are captured digitally. As a result, data have been growing at an overwhelming rate, being measured by terabytes (1012 bytes), petabytes (1015 bytes), exabytes (1018 bytes), and even by higherdimensional terms. Just think of the amount of data stored on Facebook, Twitter, or Amazon servers, or the amount of data acquired daily from scanning items at a national grocery chain such as Kroger and its affiliates. Walmart, for instance, has over one million transactions each hour, yielding more than 2.5 petabytes of data. Analytics professionals have coined the term big data to refer to massive amounts of business data from a wide variety of sources, much of which is available in real time. IBM calls these characteristics volume, variety, and v elocity. Most often, big data revolve around customer behavior and customer experiences. Big data provide an opportunity for organizations to gain a competitive advantage—if the data can be understood and analyzed effectively to make better business decisions. The volume of data continues to increase; what is considered “big” today will be even bigger tomorrow. In one study of information technology (IT) professionals in 2010, nearly half of survey respondents ranked data growth among their top three challenges. Big data are captured using sensors (for example, supermarket scanners), click streams from the Web, customer transactions, emails, tweets and social media, and other ways. Big data sets are unstructured and messy, requiring sophisticated analytics to integrate and process the data and understand the information contained in them. Because much big data are being captured in real time, they must be incorporated into business decisions at a faster rate. Processes such as fraud detection must be analyzed quickly to have value. In addition to volume, variety, and velocity, IBM proposed a fourth dimension: veracity—the level of reliability associated with data. Having highquality data and understanding the uncertainty in data are essential for good decision making. Data veracity is an important role for statistical methods. Big data can help organizations better understand and predict customer behavior and improve customer service. A study by the McKinsey Global Institute noted that, “The effective use of big data has the potential to transform economies, delivering a new wave of productivity growth and consumer surplus. Using big data will become a key basis of competition for existing companies, and will create new competitors who are able to attract employees that have the critical skills for a big data world.”26 However, understanding big data requires advanced analytics tools such as data mining and text analytics, and new technologies such as cloud computing, faster multicore processors, large memory spaces, and solidstate drives.
Data Reliability and Validity Poor data can result in poor decisions. In one situation, a distribution system design model relied on data obtained from the corporate finance department. Transportation costs were 25 Based on a presentation by Bill Franks of Teradata, “Optimizing Customer Analytics: How Customer Level Web Data Can Help,” INFORMS Conference on Business Analytics and Operations Research, April 10–12, 2011. 26 James Manyika, Michael Chui, Brad Brown, Jacques Bughin, Richard Dobbs, Charles Roxburgh, and Angela Hung Byers, “Big Data: The Next Frontier for Innovation, Competition, and Productivity,” McKinsey & Company May 2011.
Chapter 1 Introduction to Business Analytics
45
determined using a formula based on the latitude and longitude of the locations of plants and customers. But when the solution was represented on a geographic information system (GIS) mapping program, one of the customers was located in the Atlantic Ocean. Thus, data used in business decisions need to be reliable and valid. Reliability means that data are accurate and consistent. Validity means that data correctly measure what they are supposed to measure. For example, a tire pressure gauge that consistently reads several pounds of pressure below the true value is not reliable, although it is valid because it does measure tire pressure. The number of calls to a customer service desk might be counted correctly each day (and thus is a reliable measure), but it is not valid if it is used to assess customer dissatisfaction, as many calls may be simple queries. Finally, a survey question that asks a customer to rate the quality of the food in a restaurant may be neither reliable (because different customers may have conflicting perceptions) nor valid (if the intent is to measure customer satisfaction, as satisfaction generally includes other elements of service besides food).
CHECK YOUR UNDERSTANDING 1. State three examples of how data are used in different business functions. 2. How are data obtained from the Web used in marketing and business? 3. Define big data and list the four characteristics of big data. 4. Explain the concepts of data reliability and validity.
Models in Business Analytics To make an informed decision, we must be able to specify the decision alternatives that represent the choices that can be made and criteria for evaluating the alternatives. Specifying decision alternatives might be very simple; for example, you might need to choose one of three corporate health plan options. Other situations can be more complex; for example, in locating a new distribution center, it might not be possible to list just a small number of alternatives. The set of potential locations might be anywhere in the United States or even across the globe. Decision criteria might be to maximize discounted net profits, customer satisfaction, or social benefits or to minimize costs, environmental impact, or some measure of loss. Many decision problems can be formalized using a model. A model is an abstraction or representation of a real system, idea, or object. Models capture the most important features of a problem and present them in a form that is easy to interpret. A model can be as simple as a written or verbal description of some phenomenon, a visual representation such as a graph or a flowchart, or a mathematical or spreadsheet representation. Example 1.2 illustrates three ways to express a model. A decision model is a logical or mathematical representation of a problem or business situation that can be used to understand, analyze, or facilitate making a decision. Decision models can be represented in various ways, most typically with mathematical functions and spreadsheets. Spreadsheets are ideal vehicles for implementing decision models because of their versatility in managing data, evaluating different scenarios, and presenting results in a meaningful fashion. We will focus on spreadsheet models beginning with Chapter 11.
46
Chapter 1 Introduction to Business Analytics
EXAMPLE 1.2
Three Forms of a Model
Models are usually developed from theory or observation and establish relationships between actions that decision makers might take and results that they might expect, thereby allowing the decision makers to evaluate scenarios or to predict what might happen. For example, the sales of a new product, such as a firstgeneration iPad, Android phone, or 3D television, often follow a common pattern. We might represent this in one of the three following ways:
1. A simple verbal description of sales might be: The rate of sales starts small as early adopters begin to evaluate a new product and then begins to grow at an increasing rate over time as positive customer feedback spreads.
Eventually, the market begins to become saturated, and the rate of sales begins to decrease. 2. A sketch of sales as an Sshaped curve over time, as shown in Figure 1.2, is a visual model that conveys this phenomenon. 3. Finally, analysts might identify a mathematical model that characterizes this curve. Several different mathematical functions do this; one is called a Gomct pertz curve and has the formula: S = aebe , where S = sales, t = time, e is the base of natural logarithms, and a, b, and c are constants. Of course, you would not be expected to know this; that’s what analytics professionals do.
Decision models typically have three types of input: 1. Data, which are assumed to be constant for purposes of the model. Some examples are costs, machine capacities, and intercity distances. 2. Uncontrollable inputs, which are quantities that can change but cannot be directly controlled by the decision maker. Some examples are customer demand, inflation rates, and investment returns. Often, these variables are uncertain. 3. Decision options, which are controllable and can be selected at the discretion of the decision maker. Some examples are production quantities, staffing levels, and investment allocations. Decision options are often called decision variables. Decision models characterize the relationships among these inputs and the outputs of interest to the decision maker (see Figure 1.3). In this way, the user can manipulate the decision options and understand how they influence outputs, make predictions for the future, or use analytical tools to find the best decisions. Thus, decision models can be descriptive, predictive, or prescriptive and therefore are used in a wide variety of business analytics applications. ▶▶ Figure
1.2
New Product Sales over Time
Chapter 1 Introduction to Business Analytics
▶▶ Figure
1.3
The Nature of Decision Models
Inputs
47
Outputs
Data, Uncontrollable Inputs, and Decision Options
Decision Model
Measures of Performance or Behavior
Decision models complement decision makers’ intuition and often provide insights that intuition cannot. For example, one early application of analytics in marketing involved a study of sales operations. Sales representatives had to divide their time between large and small customers and between acquiring new customers and keeping old ones. The problem was to determine how the representatives should best allocate their time. Intuition suggested that they should concentrate on large customers and that it was much harder to acquire a new customer than to keep an old one. However, intuition could not tell whether they should concentrate on the 100 largest or the 1,000 largest customers, or how much effort to spend on acquiring new customers. Models of sales force effectiveness and customer response patterns provided the insight to make these decisions. However, it is important to understand that all models are only representations of the real world and, as such, cannot capture every nuance that decision makers face in reality. Decision makers must often modify the policies that models suggest to account for intangible factors that they might not have been able to incorporate into the model.
Descriptive Models Descriptive models explain behavior and allow users to evaluate potential decisions by asking “whatif?” questions. The following example illustrates a simple descriptive mathematical model. EXAMPL E 1.3
Gasoline Usage Model
Automobiles have different fuel economies (miles per gallon), and commuters drive different distances to work or school. Suppose that a state Department of Transportation (DOT) is interested in measuring the average monthly fuel consumption of commuters in a certain city. The DOT might sample a group of commuters and collect information on the number of miles driven per day, the number of driving days per month, the fuel economy of their vehicles, and additional miles driven per month for leisure and household activities. We may develop a simple descriptive model for calculating the amount of gasoline consumed, using the following symbols for the data: G m d f a
= = = = =
gallons of fuel consumed per month miles driven per day to and from work or school number of driving days per month fuel economy in miles per gallon (mpg) additional miles for leisure and household activities per month
When developing mathematical models, it is very important to use the dimensions of the variables to ensure logical consistency. In this example, we see that (m miles/day) * (d days/month) = m * d miles/month Thus, the total number of miles driven per month = m * d + a. If the vehicle gets f miles/gallon, then the total number of gallons consumed per month is
G = (m * d + a miles/month)/(f miles/gallon) = (m * d + a)/f gallons/month (1.1)
Suppose that a commuter drives 30 miles round trip to work for 20 days each month, achieves a fuel economy of 34 mpg, and drives an additional 250 miles each month. Using formula (1.1), the number of gallons consumed is G = (30 * 20 + 250)/34 = 25.0 gallons/month
48
Chapter 1 Introduction to Business Analytics
In the previous example, we have no decision options; the model is purely descriptive, but allows us to evaluate “whatif?” questions, for example, “What if we purchase a hybrid vehicle with a fuel economy of 45 miles/gallon?” “What if leisure and household activity driving increases to 400 miles/month?” Most of the models we will be using include decision options. As an example, suppose that a manufacturer has the option of producing a part in house or outsourcing it from a supplier (the decision options). Should the firm produce the part or outsource it? The decision depends on the costs of manufacturing and outsourcing, as well as the anticipated volume of demand (the uncontrollable inputs). By developing a model to evaluate the total cost of both alternatives (the outputs), the best decision can be made.
EXAMPLE 1.4 An Outsourcing Decision Model Suppose that a manufacturer can produce a part for $125/unit with a fixed cost of $50,000. The alternative is to outsource production to a supplier at a unit cost of $175. The total manufacturing and outsourcing costs can be expressed by simple mathematical formulas, where Q is the production volume:
TC (manufacturing) = $50,000 + $125 * Q (1.2) TC (outsourcing) = $175 * Q (1.3)
These formulas comprise the decision model, which simply describes what the costs of manufacturing and outsourcing are for any level of production volume. Thus, if the anticipated production volume is 1,500 units, the cost of manufacturing will be $50,000 + $125 * 1,500 = $237,500, and the cost of outsourcing would be $175 * 1,500 = $262,500;
▶▶ Figure
1.4
Graphical Illustration of BreakEven Analysis
therefore, manufacturing would be the best decision. On the other hand, if the anticipated production volume is only 800 units, the cost of manufacturing will be $50,000 + $125 * 800 = $150,000 and the cost of outsourcing would be $175 * 800 = $140,000, and the best decision would be to outsource. If we graph the two total cost formulas, we can easily see how the costs compare for different values of Q. This is shown graphically in Figure 1.4. The point at which the total costs of manufacturing and outsourcing are equal is called the breakeven volume. This can easily be found by setting TC (manufacturing) = TC (outsourcing) and solving for Q: $50,000 + $125 * Q = $175 * Q $50,000 = 50 * Q Q = 1,000
Chapter 1 Introduction to Business Analytics
49
Predictive Models Predictive models focus on what will happen in the future. Many predictive models are developed by analyzing historical data and assuming that the past is representative of the future. Example 1.5 shows how historical data might be used to develop a model that can be used to predict the impact of pricing and promotional strategies in the grocery industry.27 EXAMPL E 1.5 A Predictive SalesPromotion Model In the grocery industry, managers typically need to know how best to use pricing, coupons, and advertising strategies to influence sales. Grocers often study the relationship of sales volume to these strategies by conducting controlled experiments. That is, they implement different combinations of pricing, coupons, and advertising, observe the
sales that result, and use analytics to develop predictive models of sales as a function of these decision strategies. For example, suppose that a grocer who operates three stores in a small city varied the price, coupons (yes = 1, no = 0), and advertising expenditures in a local newspaper over a 16week period and observed the following sales:
Store1 Sales (Units)
Store 2 Sales (Units)
Store 3 Sales (Units)
0
501
510
481
0
150
772
748
775
1
0
554
528
506
6.99
1
150
838
785
834
6.49
0
0
521
519
500
6
6.49
0
150
723
790
723
7
6.49
1
0
510
556
520
8
6.49
1
150
818
773
800
Week
Price ($)
Coupon (0,1)
1
6.99
0
2
6.99
3
6.99
4 5
Advertising ($)
9
7.59
0
0
479
491
486
10
7.59
0
150
825
822
757
11
7.59
1
0
533
513
540
12
7.59
1
150
839
791
832
13
5.49
0
0
484
480
508
14
5.49
0
150
686
683
708
15
5.49
1
0
543
531
530
16
5.49
1
150
767
743
779
To better understand the relationships among price, coupons, and advertising, an analyst might have developed the following model using business analytics tools (we will see how to do this in Chapter 8): Total Sales = 1105.55 + 56.18 * Price + 123.88 * Coupon + 5.24 * Advertising (1.4)
27
In this example, the uncontrollable inputs are the sales at each store. The decision options are price, coupons, and advertising. The numerical values in the model are estimated from the data obtained from the experiment. They reflect the impact on sales of changing the decision options. For example, an increase in price of $1 results in a 56.18unit
(continued )
Roger J. Calantone, Cornelia Droge, David S. Litvack, and C. Anthony di Benedetto. “Flanking in a Price War,” Interfaces, 19, 2 (1989): 1–12.
50
Chapter 1 Introduction to Business Analytics
increase in weekly sales; using coupons (that is, setting Coupon = 1 in the model) results in a 123.88unit increase in weekly sales. The output of the model is the predicted total sales units of the product. For example, if the price is $6.99, no coupons are offered, and no advertising is done (the experiment corresponding to week 1), the model estimates sales as
We see that the actual total sales in the three stores for week 1 was 1,492. Thus, this model appears to provide good estimates for sales using the historical data. We would hope that this model would also provide good predictions of future sales. So if the grocer decides to set the price at $5.99, does not use coupons, and spends $100 in advertising, the model would predict sales to be
Total Sales = 1,105.55 + 56.18 * 6.99 + 123.88 * 0 + 5.24 * 0 = 1,498.25 units
Total Sales = 1,105.55 + 56.18 * 5.99 + 123.88 * 0 + 5.24 * 100 = 1,966.07 units
Prescriptive Models A prescriptive decision model helps decision makers to identify the best solution to a decision problem. Optimization is the process of finding a set of values for decision options that minimize or maximize some quantity of interest—profit, revenue, cost, time, and so on—called the objective function. Any set of decision options that optimizes the objective function is called an optimal solution. In a highly competitive world, where one percentage point can mean a difference of hundreds of thousands of dollars or more, knowing the best solution can mean the difference between success and failure. EXAMPLE 1.6 A Prescriptive Model for Pricing To illustrate an example of a prescriptive model, suppose that a firm wishes to determine the best pricing for one of its products to maximize revenue over the next year. A market research study has collected data that estimate the expected annual sales for different levels of pricing. Analysts determined that sales can be expressed by the following model:
Sales = 2.9485 * Price + 3,240.9
(1.5)
Because revenue equals price * sales, a model for total revenue is
Total Revenue = Price * Sales = Price * (2.9485 * Price + 3,240.9) =  2.9485 * Price2 + 3,240.9 * Price (1.6) The firm would like to identify the price that maximizes the total revenue. One way to do this would be to try different prices and search for the one that yields the highest total revenue. This would be quite tedious to do by hand or even with a calculator; however, as we will see in later chapters, spreadsheet models make this much easier.
Although the pricing model did not, most optimization models have constraints— limitations, requirements, or other restrictions that are imposed on any solution, such as “Do not exceed the allowable budget” or “Ensure that all demand is met.” For instance, a consumer products company manager would probably want to ensure that a specified level of customer service is achieved with the redesign of the distribution system. The presence of constraints makes modeling and solving optimization problems more challenging; we address constrained optimization problems later in this book, starting in Chapter 13. For some prescriptive models, analytical solutions—closedform mathematical expressions or simple formulas—can be obtained using such techniques as calculus or other types of mathematical analyses. In most cases, however, some type of computerbased procedure is needed to find an optimal solution. An algorithm is a systematic procedure that finds a solution to a problem. Researchers have developed effective algorithms to solve many types of optimization problems. For example, Microsoft Excel has a builtin addin called Solver that allows you to find optimal solutions to optimization problems formulated as spreadsheet models. We use Solver in later chapters. However, we will not be concerned with the detailed mechanics of these algorithms; our focus will be on the use of the algorithms to solve and analyze the models we develop.
Chapter 1 Introduction to Business Analytics
51
If possible, we would like to ensure that an algorithm such as the one Solver uses finds the best solution. However, some models are so complex that it is impossible to solve them optimally in a reasonable amount of computer time because of the extremely large number of computations that may be required or because they are so complex that finding the best solution cannot be guaranteed. In these cases, analysts use search algorithms—solution procedures that generally find good solutions without guarantees of finding the best one. Powerful search algorithms exist to obtain good solutions to extremely difficult optimization problems. One of these is discussed in Chapter 14.
Model Assumptions All models are based on assumptions that reflect the modeler’s view of the “real world.” Some assumptions are made to simplify the model and make it more tractable, that is, able to be easily analyzed or solved. Other assumptions might be made to better characterize historical data or past observations. The task of the modeler is to select or build an appropriate model that best represents the behavior of the real situation. For example, economic theory tells us that demand for a product is negatively related to its price. Thus, as prices increase, demand falls, and vice versa (a phenomenon that you may recognize as price elasticity—the ratio of the percentage change in demand to the percentage change in price). Different mathematical models can describe this phenomenon. In the following examples, we illustrate two of them. EXAMPL E 1.7 A Linear Demand Prediction Model A simple model to predict demand as a function of price is the linear model D = a  bP
(1.7)
where D is the demand, P is the unit price, a is a constant that estimates the demand when the price is zero, and b is the slope of the demand function. This model is most applicable when we want to predict the effect of small changes around the current price. For example, suppose we know that when the price is $100, demand is 19,000 units and that demand falls by 10 for each dollar of price increase. Using simple algebra, we can determine that a = 20,000 and b = 10. Thus, if the price is $80, the predicted demand is
▶▶ Figure
1.5
Graph of Linear Demand Model D = a  bP
D = 20,000  10(80) = 19,200 units If the price increases to $90, the model predicts demand as D = 20,000  10(90) = 19,100 units If the price is $100, demand would be D = 20,000  10(100) = 19,000 units and so on. A graph of demand as a function of price is shown in Figure 1.5 as price varies between $80 and $120. We see that there is a constant decrease in demand for each $10 increase in price, a characteristic of a linear model.
52
Chapter 1 Introduction to Business Analytics
EXAMPLE 1.8 A Nonlinear Demand Prediction Model An alternative model assumes that price elasticity is constant. In this case, the appropriate model is D = cP
d
(1.8)
where c is the demand when the price is 0 and d 7 0 is the price elasticity. To be consistent with Example 1.7, we assume that when the price is zero, demand is 20,000. Therefore, c = 20,000. We will also, as in Example 1.7, assume that when the price is $100, D = 19,000. Using these values in equation (1.8), we can determine the value for d as 0.0111382 (we can do this mathematically using logarithms, but we’ll see how to do this very easily using Excel in Chapter 11). Thus, if the price is $80, then the predicted demand is
If the price is 90, the demand would be D = 20,000(90)  0.0111382 = 19,022 If the price is 100, demand is D = 20,000(100)  0.0111382 = 19,000 A graph of demand as a function of price is shown in Figure 1.6. The predicted demand falls in a slight nonlinear fashion as price increases. For example, demand decreases by 25 units when the price increases from $80 to $90, but only by 22 units when the price increases from $90 to $100. If the price increases to $110, you would see a smaller decrease in demand. Therefore, we see a nonlinear relationship, in contrast to Example 1.7.
D = 20,000(80)  0.0111382 = 19,047
Both models in Examples 1.7 and 1.8 make different predictions of demand for different prices (other than $90). Which model is best? The answer may be neither. First of all, the development of realistic models requires many price point changes within a carefully designed experiment. Second, it should also include data on competition and customer disposable income, both of which are hard to determine. Nevertheless, it is possible to develop price elasticity models with limited price ranges and narrow customer segments. A good starting point would be to create a historical database with detailed information on all past pricing actions. Unfortunately, practitioners have observed that such models are not widely used in retail marketing, suggesting ample opportunity to apply business analytics.28 ▶▶ Figure
1.6
Graph of Nonlinear Demand Model D = cP d
28
Zhang, Clay Duan, and Arun Muthupalaniappan, “Analytics Applications in Consumer Credit and Retail Marketing,” Analytics Magazine (November  December 2011): 27–33.
Chapter 1 Introduction to Business Analytics
53
Uncertainty and Risk As we all know, the future is always uncertain. Thus, many predictive models incorporate uncertainty and help decision makers analyze the risks associated with their decisions. Uncertainty is imperfect knowledge of what will happen; risk is associated with the consequences and likelihood of what might happen. For example, the change in the stock price of Apple on the next day of trading is uncertain. If you own Apple stock, then you face the risk of losing money if the stock price falls. If you don’t own any stock, the price is still uncertain, although you would not have any risk. Risk is evaluated by the magnitude of the consequences and the likelihood that they would occur. For example, a 10% drop in a stock price would incur a higher risk if you own $1 million worth of that stock than if you only owned $1,000 worth of that stock. Similarly, if the chances of a 10% drop were 1 in 5, the risk would be higher than if the chances were only 1 in 100. The importance of risk in business has long been recognized. The renowned management writer Peter Drucker observed in 1974: To try to eliminate risk in business enterprise is futile. Risk is inherent in the commitment of present resources to future expectations. Indeed, economic progress can be defined as the ability to take greater risks. The attempt to eliminate risks, even the attempt to minimize them, can only make them irrational and unbearable. It can only result in the greatest risk of all: rigidity.29 Consideration of risk is a vital element of decision making. For instance, you would probably not choose an investment simply on the basis of the return you might expect because, typically, higher returns are associated with higher risk. Therefore, you have to make a tradeoff between the benefits of greater rewards and the risks of potential losses. Analytic models can help assess risk. A model in which some of the model input information is uncertain is often called a stochastic, or probabilistic, model. In contrast, a deterministic model is one in which all model input information is either known or assumed to be known with certainty. For instance, suppose that customer demand is an important element of some model. We can make the assumption that the demand is known with certainty; say, 5,000 units per month. In this case, we would be dealing with a deterministic model. On the other hand, suppose we have evidence to indicate that demand is uncertain, with an average value of 5,000 units per month, but which typically varies between 3,200 and 6,800 units. If we make this assumption, we would be dealing with a stochastic model. Stochastic models are useful in analyzing uncertainty in realworld situations, and we will discuss these later in this book. CHECK YOUR UNDERSTANDING 1. Define a model and state three common forms of a model. 2. Explain the elements of a decision model. 3. Explain how decision models are used for descriptive, predictive, and prescriptive applications. 4. Define optimization and the characteristics of optimization models. 5. Explain the importance of assumptions in building decision models. 6. What is the difference between uncertainty and risk? 29
P. F. Drucker, The Manager and the Management Sciences in Management: Tasks, Responsibilities, Practices (London: Harper and Row, 1974).
54
Chapter 1 Introduction to Business Analytics
Problem Solving with Analytics The fundamental purpose of analytics is to help managers solve problems and make decisions. The techniques of analytics represent only a portion of the overall problemsolving and decisionmaking process. Problem solving is the activity associated with defining, analyzing, and solving a problem and selecting an appropriate solution that solves a problem. Problem solving consists of several phases: 1. 2. 3. 4. 5. 6.
Recognizing a problem Defining the problem Structuring the problem Analyzing the problem Interpreting results and making a decision Implementing the solution
Recognizing a Problem Managers at different organizational levels face different types of problems. In a manufacturing firm, for instance, top managers face decisions regarding allocating financial resources, building or expanding facilities, determining product mix, and strategically sourcing production. Middle managers in operations develop distribution plans, production and inventory schedules, and staffing plans. Finance managers analyze risks, determine investment strategies, and make pricing decisions. Marketing managers develop advertising plans and make sales force allocation decisions. In manufacturing operations, problems involve the size of daily production runs, individual machine schedules, and worker assignments. Whatever the problem, the first step is to realize that it exists. How are problems recognized? Problems exist when there is a gap between what is happening and what we think should be happening. For example, a consumer products manager might feel that distribution costs are too high. This recognition might result from comparing performance with that of a competitor, or observing an increasing trend compared to previous years.
Defining the Problem The second step in the problemsolving process is to clearly define the problem. Finding the real problem and distinguishing it from symptoms that are observed is a critical step. For example, high distribution costs might stem from inefficiencies in routing trucks, poor location of distribution centers, or external factors such as increasing fuel costs. The problem might be defined as improving the routing process, redesigning the entire distribution system, or optimally hedging fuel purchases. Defining problems is not a trivial task. The complexity of a problem increases when the following occur: ■■ ■■ ■■ ■■ ■■ ■■
The number of potential courses of action is large. The problem belongs to a group rather than to an individual. The problem solver has several competing objectives. External groups or individuals are affected by the problem. The problem solver and the true owner of the problem—the person who experiences the problem and is responsible for getting it solved—are not the same. Time limitations are important.
Chapter 1 Introduction to Business Analytics
55
These factors make it difficult to develop meaningful objectives and characterize the range of potential decisions. In defining problems, it is important to involve all people who make the decisions or who may be affected by them.
Structuring the Problem This usually involves stating goals and objectives, characterizing the possible decisions, and identifying any constraints or restrictions. For example, if the problem is to redesign a distribution system, decisions might involve new locations for manufacturing plants and warehouses (where?), new assignments of products to plants (which ones?), and the amount of each product to ship from different warehouses to customers (how much?). The goal of cost reduction might be measured by the total delivered cost of the product. The manager would probably want to ensure that a specified level of customer service—for instance, being able to deliver orders within 48 hours—is achieved with the redesign. This is an example of a constraint. Structuring a problem often involves developing a formal model.
Analyzing the Problem Here is where analytics plays a major role. Analysis involves some sort of experimentation or solution process, such as evaluating different scenarios, analyzing risks associated with various decision alternatives, finding a solution that meets certain goals, or determining an optimal solution. Analytics professionals have spent decades developing and refining a variety of approaches to address different types of problems. Much of this book is devoted to helping you understand these techniques and gain a basic facility in using them.
Interpreting Results and Making a Decision Interpreting the results from the analysis phase is crucial in making good decisions. Models cannot capture every detail of the real problem, and managers must understand the limitations of models and their underlying assumptions and often incorporate judgment into making a decision. For example, in locating a facility, we might use an analytical procedure to find a “central” location; however, many other considerations must be included in the decision, such as highway access, labor supply, and facility cost. Thus, the location specified by an analytical solution might not be the exact location the company actually chooses.
Implementing the Solution This simply means making the solution work in the organization, or translating the results of a model back to the real world. This generally requires providing adequate resources, motivating employees, eliminating resistance to change, modifying organizational policies, and developing trust. Problems and their solutions affect people: customers, suppliers, and employees. All must be an important part of the problemsolving process. Sensitivity to political and organizational issues is an important skill that managers and analytical professionals alike must possess when solving problems. In each of these steps, good communication is vital. Analytics professionals need to be able to communicate with managers and clients to understand the business context of the problem and be able to explain results clearly and effectively. Such skills as constructing good visual charts and spreadsheets that are easy to understand are vital to users of analytics. We emphasize these skills throughout this book.
56
Chapter 1 Introduction to Business Analytics
ANALYTICS IN PRACTICE: Developing Effective Analytical Tools
at HewlettPackard30
HewlettPackard (HP) uses analytics extensively. Many applications are used by managers with little knowledge of analytics. These require that analytical tools be easily understood. Based on years of experience, HP analysts compiled some key lessons. Before creating an analytical decision tool, HP asks three questions:
hans12/Fotolia
1. Will analytics solve the problem? Will the tool enable a better solution? Should other, nonanalytical solutions be used? Are there organizational or other issues that must be resolved? Often, what may appear to be an analytical problem may actually be rooted in problems of incentive misalignment, unclear ownership and accountability, or business strategy. 2. Can we leverage an existing solution? Before “reinventing the wheel,” can existing solutions address the problem? What are the costs and benefits? 3. Is a decision model really needed? Can simple decision guidelines be used instead of a formal decision tool? Once a decision is made to develop an analytical tool, several guidelines are used to increase the chances of successful implementation: ■■
■■
■■
Use prototyping, a quick working version of the tool designed to test its features and gather feedback. Build insight, not black boxes. A “black box” tool is one that generates an answer, but may not provide confidence to the user. Interactive tools that create insights to support a decision provide better information. Remove unneeded complexity. Simpler is better. A good tool can be used without expert support.
■■
■■
Partner with end users in discovery and design. Decision makers who will actually use the tool should be involved in its development. Develop an analytic champion. Someone (ideally, the actual decision maker) who is knowledgeable about the solution and close to it must champion the process.
CHECK YOUR UNDERSTANDING 1. List the major phases of problem solving and explain each. 2. What lessons did HewlettPackard learn about using analytics?
30
Based on Thomas Olavson and Chris Fry, “Spreadsheet DecisionSupport Tools: Lessons Learned at HewlettPackard,” Interfaces, 38, 4, July–August 2008: 300–310.
Chapter 1 Introduction to Business Analytics
57
KEY TERMS Algorithm Big data Business analytics (analytics) Business intelligence (BI) Constraint Data mining Decision model Decision options (decision variables) Decision support systems (DSS) Descriptive analytics Deterministic model Information systems (IS) Model Modeling and optimization Objective function Operations Research/Management Science (OR/MS)
Optimal solution Optimization Predictive analytics Prescriptive analytics Price elasticity Problem solving Reliability Risk Search algorithm Simulation and risk analysis Statistics Stochastic (probabilistic) model Tag cloud Uncertainty Validity Visualization Whatif analysis
CHAPTER 1 TECHNOLOGY HELP Useful Excel Functions (see Appendix A1) MIN(range) Finds the smallest value in a range of cells. MAX(range) Finds the largest value in a range of cells. SUM(range) Finds the sum of values in a range of cells. AVERAGE(range) Finds the average of the values in a range of cells. COUNT(range) Finds the number of cells in a range that contain numbers.
COUNTIF(range, criteria) Finds the number of cells within a range that meet a specified criterion. NPV(rate, value1, value2, . . . ) Calculates the net present value of an investment by using a discount rate and a series of future payments (negative values) and income (positive values). DATEDIF(startdate, enddate, time unit) Computes the number of whole years, months, or days between two dates.
PROBLEMS AND EXERCISES What Is Business Analytics? 1. Discuss the use of business analytics in sports, such
as tennis, cricket, swimming, and football. Identify as many opportunities as you can for each. 2. How might analytics be used in the following situations? a. Planning a business b. Understanding customer behavior c. Solving human resource problems
3. Swissotel, London, is working on improving cus
tomer satisfaction and engagement. At the same time, Swissotel wants to ensure that the staff members are also satisfied. It has noticed that there are several areas that will need to be analyzed. The problem facing the hotel management is identifying what information is necessary to analyze a situation and improve customer and staff satisfaction levels. What data would Swissotel need to collect from its guests to facilitate good decisions? How might this data and business analytics help the hotel management?
58
Chapter 1 Introduction to Business Analytics
Descriptive, Predictive, and Prescriptive Analytics 4. For each of the following scenarios, state whether
descriptive, predictive, or prescriptive analytics tools would most likely be used. a. An insurance firm wants to analyze fraudulent claims to predict their future losses. b. A store manager has to review a store’s historical operations, sales, financials, and customer reports to improve the store’s business. c. The Met Office would like to forecast weather for nine to 10 days more accurately than the 24hour forecasts, compared to the forecasting done 40 years ago. d. Autonomous or selfdriving car programmers have to make millions of calculations for each trip to determine directions and speed. e. A secretary needs to summarize past events such
as regional sales, customer attrition, and success of marketing campaigns for her manager to understand their company’s progression. f. A fuel producer wants to identify the factors affecting the price of oil and gas to get the best terms and hedge risks. g. A mechanical engineer needs to predict the failure rate of the equipment in the next 6 months to help reduce maintenance costs and improve power availability. h. A travel agency is setting the price of their travel packages to attract more customers and wants to know the trends in popular travel destinations. Models in Business Analytics 5. A firm installs 1,500 air conditioners that need to be
checked every six months. The firm can hire a team from its logistics department at a fixed cost of €6,000. This team will check each unit for €15.00. The firm can also outsource this at a cost of €17.00 inclusive of all charges. a. For the given number of units, compute the firm’s total cost of checking for both options. Which is a better decision? b. Find the breakeven volume and characterize the range of volumes for which it is more economical to outsource. 6. Use the model developed in Example 1.5 to pre
dict the total sales for weeks 2 through 16, and
compare the results to the observed sales. Does the accuracy of the model seem to be different when coupons are used or not? When advertising is used or not? 7. The yearly government tax revenue (in $ millions)
depends on customer income, business profits, and capital gains defined as tax = 2.556 + 3.799 * income + 1.583 * profit + 0.739 * capital. a. Interpret the numbers in this model. b. What is the government tax revenue if a customer earns $1.2 million, the business’ profits amount to $8.63 million, and the capital gains are $5.76 million? 8. In a toy manufacturing company, the manufacturer
needs to pay rent (p) each month and monthly electricity (q). To produce each toy, the company incurs costs for plastic (r) and for cloth (s). The variables p, q, r, and s are positive constants. Given the cost model as cost = fixed cost + variable cost a. Define the cost model with the given unknowns. b. How does plastic and cloth affect the cost model? c. Will the variable plastic influence the cloth variable? 9. An insurance company is estimating the risk model
based on customer age is define as R = aebx where a is a fixed constant and b is a constant that depends on the customer age. They categorize the age into 5 categories as age 6 10 with b = 0.05, 10 6 age … 20 with b = 0.02, 20 6 age … 30 with b = 0.01, 30 6 age … 40 with b = 0.03 and age 7 40 with b = 0.08. Sketch the graphs for each model and interpret them. Are the graphs reasonable? 10. A tablet manufacturer is preparing to set the price of
a new model. It examines the demand and cost structure, and determines the following model to represent the relationship between demand and price D = 50 + 2P The finance department estimates that the total costs is represented by C = 100 + 0.3D Develop a model for the total profit in terms of the price, P.
Chapter 1 Introduction to Business Analytics
was able to estimate the regions of the aircraft most likely to be hit by enemy gunfire, with the objective of reinforcing these regions with special armor. What difficulties do you see here?
Problem Solving with Analytics 11. In this chapter, we noted the importance of defining
and analyzing a problem prior to attempting to find a solution. Consider this example: One of the earliest operations research groups during World War II was conducting a study on the optimum utilization of Spitfire and Hurricane aircraft during the Battle of Britain. Whenever one of these planes returned from battle, the locations of the bullet holes on it were carefully plotted. By repeatedly recording these data over time, and studying the clusters of data, the group
59
12. A parttime professor of a business program at a
university in Germany always tells her students that one of the first questions they should ask their new employers is whether the organization has a good customer program. Why do you think the professor encourages her students to ask this specific question?
CASE: PERFORMANCE LAWN EQUIPMENT In each chapter of this book, we use a a fictitious company, Performance Lawn Equipment (PLE), within a case exercise for applying the tools and techniques introduced in the chapter.31 To put the case in perspective, we first provide some background about the company, so that the applications of business analytic tools will be more meaningful. PLE, headquartered in St. Louis, Missouri, is a privately owned designer and producer of traditional lawn mowers used by homeowners. In the past ten years, PLE has added another key product, a mediumsize diesel power lawn tractor with front and rear power takeoffs, Class I threepoint hitches, fourwheel drive, power steering, and full hydraulics. This equipment is built primarily for a niche market consisting of large estates, including golf and country clubs, resorts, private estates, city parks, large commercial complexes, lawn care service providers, private homeowners with five or more acres, and government (federal, state, and local) parks, building complexes, and military bases. PLE provides most of the products to dealerships, which, in turn, sell directly to end users. PLE employs 1,660 people worldwide. About half the workforce is based in St. Louis; the remainder is split among their manufacturing plants. In the United States, the focus of sales is on the eastern seaboard, California, the Southeast, and the south central states, which have the greatest concentration of customers. Outside the United States, PLE’s sales include a European market, a growing South American market, and developing markets in the Pacific Rim and China. Both end users and dealers have been established as important customers for PLE. Collection and analysis of enduser data showed that satisfaction with the products depends on high quality, easy attachment/dismount of implements, low maintenance, price value, and service. 31
For dealers, key requirements are high quality, parts and feature availability, rapid restock, discounts, and timeliness of support. PLE has several key suppliers: Mitsitsiu, Inc., the sole source of all diesel engines; LANTO Axles, Inc., which provides tractor axles; Schorst Fabrication, which provides subassemblies; Cuberillo, Inc, supplier of transmissions; and Specialty Machining, Inc., a supplier of precision machine parts. To help manage the company, PLE managers have developed a “balanced scorecard” of measures. These data, which are summarized shortly, are stored in the form of a Microsoft Excel workbook (Performance Lawn Equipment) accompanying this book. The database contains various measures captured on a monthly or quarterly basis and is used by various managers to evaluate business performance. Data for each of the key measures are stored in a separate worksheet. A summary of these worksheets is given next: ■■
■■
■■
Dealer Satisfaction, measured on a scale of 1–5 (1 = poor, 2 = less than average, 3 = average, 4 = above average, and 5 = excellent). Each year, dealers in each region are surveyed about their overall satisfaction with PLE. The worksheet contains summary data from surveys for the past five years. EndUser Satisfaction, measured on the same scale as dealers. Each year, 100 users from each region are surveyed. The worksheet contains summary data for the past five years. Customer Survey, results from a survey for customer ratings of specific attributes of PLE tractors: quality, ease of use, price, and service on the same 1–5 scale. This sheet contains 200 observations of customer ratings.
The case scenario was based on Gateway Estate Lawn Equipment Co. Case Study, used for the 1997 Malcolm Baldrige National Quality Award Examiner Training course. This material is in the public domain. The database, however, was developed by the author.
60 ■■
■■
■■
■■
■■
■■
■■
■■
■■
■■
Chapter 1 Introduction to Business Analytics
Complaints, which shows the number of complaints registered by all customers each month in each of PLE’s five regions (North America, South America, Europe, the Pacific, and China). Mower Unit Sales and Tractor Unit Sales, which provide sales by product by region on a monthly basis. Unit sales for each region are aggregated to obtain world sales figures. Industry Mower Total Sales and Industry Tractor Total Sales, which list the number of units sold by all producers by region. Unit Production Costs, which provides monthly accounting estimates of the variable cost per unit for manufacturing tractors and mowers over the past five years. Operating and Interest Expenses, which provides monthly administrative, depreciation, and interest expenses at the corporate level. OnTime Delivery, which provides the number of deliveries made each month from each of PLE’s major suppliers, the number on time, and the percent on time. Defects After Delivery, which shows the number of defects in supplierprovided material found in all shipments received from suppliers. Time to Pay Suppliers, which provides measurements in days from the time the invoice is received until payment is sent. Response Time, which gives samples of the times taken by PLE customerservice personnel to respond to service calls by quarter over the past two years. Employee Satisfaction, which provides data for the past four years of internal surveys of employees to determine their overall satisfaction with their jobs, using the same scale used for customers. Employees are surveyed quarterly, and results are stratified by employee category: design and production, managerial, and sales/administrative support.
In addition to these business measures, the PLE database contains worksheets with data from special studies: ■■ ■■
■■
■■
Engines, which lists 50 samples of the time required to produce a lawn mower blade using a new technology. Transmission Costs, which provides the results of 30 samples each for the current process used to produce tractor transmissions and two proposed new processes. Blade Weight, which provides samples of mower blade weights to evaluate the consistency of the production process. Mower Test, which lists test results of mower functional performance after assembly for 30 samples of 100 units each.
■■
■■
■■
■■
Employee Retention, data from a study of employee duration (length of hire) with PLE. The 40 subjects were identified by reviewing hires from ten years prior and identifying those who were involved in managerial positions (either hired into management or promoted into management) at some time in this tenyear period. Shipping Cost, which gives the unit shipping cost for mowers and tractors from existing and proposed plants for a supply chain design study. Fixed Cost, which lists the fixed cost to expand existing plants or build new facilities, also as part of the supply chain design study. Purchasing Survey, which provides data obtained from a thirdparty survey of purchasing managers of customers of Performance Lawn Care.
Elizabeth Burke has recently joined the PLE management team to oversee production operations. She has reviewed the types of data that the company collects and has assigned you the responsibility to be her chief analyst in the coming weeks. She has asked you to do some preliminary analysis of the data for the company. 1. First, she would like you to edit the worksheets Dealer Satisfaction and EndUser Satisfaction to display the total number of responses to each level of the survey scale across all regions for each year. 2. Second, she wants a count of the number of failures in the worksheet Mower Test. 3. Next, Elizabeth has provided you with prices for PLE products for the past five years: Year
Mower Price
Tractor Price
2014
$150
$3,250
2015
$175
$3,400
2016
$180
$3,600
2017
$185
$3,700
2018
$190
$3,800
Create a new worksheet to compute gross revenues by month and region, as well as worldwide totals, for each product using the data in Mower Unit Sales and Tractor Unit Sales. 4. Finally, she wants to know the market share for each product and region by month based on the PLE and industry sales data, and the average market share by region over the five years. Summarize all your findings in a report to Ms. Burke.
APPENDIX
Basic Excel Skills
A1
To be able to apply the procedures and techniques that you will learn in this book, it is necessary for you to be relatively proficient in using Excel. We assume that you are familiar with the most elementary spreadsheet concepts and procedures, such as ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■
opening, saving, and printing files; using workbooks and worksheets; moving around a spreadsheet; selecting cells and ranges; inserting/deleting rows and columns; entering and editing text, numerical data, and formulas in cells; formatting data (number, currency, decimal places, etc.); working with text strings; formatting data and text; and modifying the appearance of the spreadsheet using borders, shading, and so on.
Menus and commands in Excel reside in the “ribbon” shown in Figure A1.1. All Excel discussions in this book will be based on Excel 2016 for Windows; if you use Excel 2016 for Mac, some differences may exist, and we may point these out as appropriate. Menus and commands are arranged in logical groups under different tabs (Home, Insert, Formulas, and so on); small triangles pointing downward indicate menus of additional choices. We often refer to certain commands or options and where they may be found in the ribbon. For instance, in the Mac version, groups are not specified. Excel provides an addin called the Analysis Toolpak, which contains a variety of tools for statistical computation, and Solver, which is used for optimization. They will be found in the Data tab ribbon; you should ensure that these are activated. To activate them in Windows, click the File tab and then Options in the left column. Choose AddIns from the left column. At the bottom of the dialog, make sure Excel Addins is selected in the Manage: box and click Go. In the AddIns dialog, if Analysis Toolpak, Analysis Toolpak VBA, and Solver Addin are not checked, simply check the boxes and click OK. You will not have to repeat this procedure every time you run Excel in the future. On Excel 2016 for Mac, go to Tools 7 Excel Addins and select both Analysis Toolpak and Solver.
▲▲ Figure
A1.1
Excel Ribbons for Windows and Mac 61
62
Appendix A1 Basic Excel Skills
Excel Formulas and Addressing Formulas in Excel use common mathematical operators: ■■ ■■ ■■ ■■
addition (+) subtraction () multiplication (*) division (/)
Exponentiation uses the ^ symbol; for example, 25 is written as 2^5 in an Excel formula. Cell references in formulas can be written either with relative addresses or absolute addresses. A relative address uses just the row and column label in the cell reference (for example, A4 or C21); an absolute address uses a dollar sign ($ sign) before either the row or column label or both (for example, $A2, C$21, or $B$15). Which one we choose makes a critical difference if you copy the cell formulas. If only relative addressing is used, then copying a formula to another cell changes the cell references by the number of rows or columns in the direction that the formula is copied. So, for instance, if we would use a formula in cell B8, =B4B5*A8, and copy it to cell C9 (one column to the right and one row down), all the cell references are increased by one and the formula would be changed to =C5C6*B9. Using a $ sign before a row label (for example, B$4) keeps the reference fixed to row 4 but allows the column reference to change if the formula is copied to another cell. Similarly, using a $ sign before a column label (for example, $B4) keeps the reference to column B fixed but allows the row reference to change. Finally, using a $ sign before both the row and column labels (for example, $B$4) keeps the reference to cell B4 fixed no EXAMPLE A1.1
Implementing PriceDemand Models in Excel
In Chapter 1, we described two models for predicting demand as a function of price: D = a  bP
=$B$4  $B$5*A8
D = cP  d
To calculate the demand in cell E8 for the nonlinear model, we use the formula
and Figure A1.2 shows a spreadsheet (Excel file Demand Prediction Models) for calculating demand for different prices using each of these models. For example, to
▶▶ Figure
A1.2
Excel Models for Demand Prediction
alculate the demand in cell B8 for the linear model, we c use the formula
=$E$4*D8^  $E$5 Note how the absolute addresses are used so that as these formulas are copied down, the demand is computed correctly.
Appendix A1 Basic Excel Skills
63
matter where the formula is copied. You should be very careful to use relative and absolute addressing appropriately in your models, especially when copying formulas.
Copying Formulas Excel provides several ways of copying formulas to different cells. This is extremely useful in building decision models, because many models require replication of formulas for different periods of time, similar products, and so on. The easiest way is to select the cell with the formula to be copied, and then press CtrlC on your Windows keyboard or CommandC on a Mac, click on the cell you wish to copy to, and then press CtrlV in Windows or CommandV on a Mac. You may also enter a formula directly in a range of cells without copying and pasting by selecting the range, typing in the formula, and pressing CtrlEnter in Windows or CommandEnter on a Mac. To copy a formula from a single cell or range of cells down a column or across a row, first select the cell or range, click and hold the mouse on the small square in the lower righthand corner of the cell (the “fill handle”), and drag the formula to the “target” cells to which you wish to copy.
Useful Excel Tips ■■
■■
■■
■■
■■
Split Screen. You may split the worksheet horizontally and/or vertically to view different parts of the worksheet at the same time. The vertical splitter bar is just to the right of the bottom scroll bar, and the horizontal splitter bar is just above the righthand scroll bar. Position your cursor over one of these until it changes shape, click, and drag the splitter bar to the left or down. Column and Row Widths. Many times a cell contains a number that is too large to display properly because the column width is too small. You may change the column width to fit the largest value or text string anywhere in the column by positioning the cursor to the right of the column label so that it changes to a cross with horizontal arrows and then doubleclicking. You may also move the arrow to the left or right to manually change the column width. You may change the row heights in a similar fashion by moving the cursor below the row number label. This can be especially useful if you have a very long formula to display. To break a formula within a cell, position the cursor at the break point in the formula bar and press AltEnter. Displaying Formulas in Worksheets. Choose Show Formulas in the Formulas tab. You may also press Ctrl ~ in either Windows or Mac to toggle formulas on and off. You often need to change the column width to display the formulas properly. Displaying Grid Lines and Row and Column Headers for Printing. Check the Print boxes for gridlines and headings in the Sheet Options group under the Page Layout tab. Note that the Print command can be found by clicking on the Office button in Windows or under the File menu in Mac. Filling a Range with a Series of Numbers. Suppose you want to build a worksheet for entering 100 data values. It would be tedious to have to enter the numbers from 1 to 100 one at a time. Simply fill in the first few values in the series and highlight them. Then click and drag the small square (fill handle) in the lower righthand corner down (Excel will show a small popup window that tells you the last value in the range) until you have filled in the column to 100; then release the mouse.
64
Appendix A1 Basic Excel Skills
Excel Functions Functions are used to perform special calculations in cells and are used extensively in business analytics applications. All Excel functions require an equal sign and a function name followed by parentheses, in which you specify arguments for the function.
Basic Excel Functions Some of the more common functions that we will use in applications include the following: MIN(range)—finds the smallest value in a range of cells MAX(range)—finds the largest value in a range of cells SUM(range)—finds the sum of values in a range of cells AVERAGE(range)—finds the average of the values in a range of cells COUNT(range)—finds the number of cells in a range that contain numbers COUNTIF(range, criteria)—finds the number of cells within a range that meet a specified criterion Logical functions, such as IF, AND, OR, and VLOOKUP will be discussed in Chapter 2. The COUNTIF function counts the number of cells within a range that meet a criterion you specify. For example, you can count all the cells that start with a certain letter, or you can count all the cells that contain a number that is larger or smaller than a number you specify. Examples of criteria are 100, “7100”, a cell reference such as A4, and a text string such as “Facebook.” Note that text and logical formulas must be enclosed in quotes. See Excel Help for other examples. Excel has other useful COUNTtype functions: COUNTA counts the number of nonblank cells in a range, and COUNTBLANK counts the number of blank cells in a range. In addition, COUNTIFS(range1, criterion1, range2, criterion2, . . . , range_n, criterion_n) finds the number of cells within multiple ranges that meet specific criteria for each range. We illustrate these functions using the Purchase Orders data set in Example A1.2. EXAMPLE A1.2
Using Basic Excel Functions
In the Purchase Orders data set, we will find the following: ■■ ■■ ■■ ■■ ■■ ■■
■■ ■■ ■■
Smallest and largest quantity of any item ordered Total order costs Average number of months per order for accounts payable Number of purchase orders placed Number of orders placed for Orings Number of orders with A/P (accounts payable) terms shorter than 30 months Number of Oring orders from Spacetime Technologies Total cost of all airframe fasteners Total cost of airframe fasteners purchased from Alum Sheeting
The results are shown in Figure A1.3. In this figure, we used the splitscreen feature in Excel to reduce the number of rows shown in the spreadsheet. To find the smallest and largest quantity of any item ordered, we use the MIN and MAX functions for the data in column F. Thus, the formula in cell B99 is = MIN(F4:F97) and the formula in cell B100 is = MAX(F4:F97). To find the total order costs,
we sum the data in column G using the SUM function: = SUM(G4:G97); this is the formula in cell B101. To find the average number of A/P months, we use the AVERAGE function for the data in column H. The formula in cell B102 is = AVERAGE(H4:H97). To find the number of purchase orders placed, use the COUNT function. Note that the COUNT function counts only the number of cells in a range that contain numbers, so we could not use it in columns A, B, or D; however, any other column would be acceptable. Using the item numbers in column C, the formula in cell B103 is = COUNT(C4:C97). To find the number of orders placed for Orings, we use the COUNTIF function. For this example, the formula used in cell B104 is = COUNTIF(D4:D97, “[email protected]”). We could have also used the cell reference for any cell containing the text ORing, such as = COUNTIF(D4:D97, D12). To find the number of orders with A/P terms less than 30 months, we use the formula = COUNTIF(H4:H97, “630”) in cell B105. Finally, to count the number of Oring orders for Spacetime Technologies, we use = COUNTIFS(D4:D97, “ORing”, A4:A97, “Spacetime Technologies”).
Appendix A1 Basic Excel Skills
65
IFtype functions are also available for other calculations. For example, the functions SUMIF, AVERAGEIF, SUMIFS, and AVERAGEIFS can be used to embed IF logic within mathematical functions. For instance, the syntax of SUMIF is SUMIF(range, criterion, [sum range]); sum range is an optional argument that allows you to add cells in a different range. Thus, in the Purchase Orders database, to find the total cost of all airframe fasteners, we would use =SUMIF(D4:D97, “Airframe fasteners”, G4:G97) This function looks for airframe fasteners in the range D4:D97, but then sums the associated values in column G (cost per order). The arguments for SUMIFS and AVERAGEIFS are (sumrange, range1, criterion1, range2, criterion2, . . . ,rangeN, criterionN). For example, the function =SUMIFS(F4:F97,A4:A97,“Alum Sheeting”, D4:D97, “Airframe fasteners”) will find the total quantity (from the sumrange in column F) of all airframe fasteners purchased from Alum Sheeting.
Functions for Specific Applications Excel has a wide variety of other functions for statistical, financial, and other applications, many of which we introduce and use throughout the text. For instance, some financial models that we develop require the calculation of net present value (NPV). Net present value (also called discounted cash flow) measures the worth of a stream of cash flows, taking into account the time value of money. That is, a cash flow of F dollars t time periods in the future is worth F> 11 + i2 t dollars today, where i is the discount rate. The discount rate reflects the opportunity costs of spending funds now versus achieving a return through another investment, as well as the risks associated with not receiving returns until a later
▲▲ Figure
A1.3
Application of Basic Excel Functions to Purchase Orders Data
66
Appendix A1 Basic Excel Skills
time. The sum of the present values of all cash flows over a stated time horizon is the net present value:
n Ft NPV = a t (A1.1) t = 0 11 + i2
where Ft = cash flow in period t. A positive NPV means that the investment will provide added value because the projected return exceeds the discount rate. The Excel function NPV(rate, value1, value2, . . .) calculates the net present value of an investment by using a discount rate and a series of future payments (negative values) and income (positive values). Rate is the value of the discount rate i over the length of one period, and value1, value2, . . . are 1 to 29 arguments representing the payments and income for each period. The values must be equally spaced in time and are assumed to occur at the end of each period. The NPV investment begins one period before the date of the value1 cash flow and ends with the last cash flow in the list. The NPV calculation is based on future cash flows. If the first cash flow (such as an initial investment or fixed cost) occurs at the beginning of the first period, then it must be added to the NPV result and not included in the function arguments. EXAMPLE A1.3
Using the NPV Function
A company is introducing a new product. The fixed cost for marketing and distribution is $25,000 and is incurred just prior to launch. The forecasted net sales revenues for the first six months are shown in Figure A1.4. The formula in
cell B8 computes the net present value of these cash flows as = NPV(B6, C4:H4)B5. Note that the fixed cost is not a future cash flow and is not included in the NPV function arguments.
Insert Function The easiest way to locate a particular function is to select a cell and click on the Insert Function button 3fx 4, which can be found under the ribbon next to the formula bar and also in the Function Library group in the Formulas tab. You may either type in a description in the search field, such as “net present value,” or select a category, such as “financial,” from the dropdown box. This feature is particularly useful if you know what function to use but are not sure of what arguments to enter because it will guide you in entering the appropriate data for the function arguments. Figure A1.5 shows the dialog from which you may select the function you wish to use. For example, if we would choose the COUNTIF function, the dialog in Figure A1.6 appears. When you click in an input cell, a description of the argument is shown. Thus, if you are not sure what to enter for the range, the explanation in Figure A1.6 ▶▶ Figure
A1.4
Net Present Value Calculation
▶▶ Figure
Appendix A1 Basic Excel Skills
67
A1.5
Insert Function Dialog
▶▶ Figure
A1.6
Function Arguments Dialog for COUNTIF
will help you. For further information, you could click on the Help button in the lower lefthand corner.
Date and Time Functions In many analytics applications, a database might contain dates, such as when an order is placed or when an employee was hired. Excel can display a date in a variety of formats, such as 2/14/17 or 14Feb17. You may choose the standard date format (for example, 2/14/17) by selecting Date in the Number formatting box or select a custom format by selecting C ustom in the Number box. Excel stores dates in a serial format. January 1, 1900 is day 1, and each subsequent day is numbered sequentially. Both the current date and January 1, 1900 are included in the count. So 2/14/17 is 42,780 when expressed in this format. This means there are 42,780 days between January 1, 1900 and February 14, 2017 (including both days). Therefore, to determine the number of days between two dates, you can simply subtract them. Another useful date function is DATEDIF (which surprisingly doesn’t appear in the Insert Function list!), which can compute the number of whole years, months, or days between two dates. The syntax of DATEDIF is DATEDIF(startdate, enddate, time unit) The time unit can be “y,” “m,” or “d.” For instance, DATEDIF(4/26/89, 2/14/17, “y”) will return 27 (years), while DATEDIF(4/26/89, 2/14/17, “m”) will return 333 (months).
68
Appendix A1 Basic Excel Skills
EXAMPLE A1.4
Computing Lead Times
In the Purchase Orders database, we will compute the lead time for each order, that is, the number of days between the order date and arrival date. Figure A1.7 shows the use of
the DATEDIF function. Alternatively, we could have simply subtracted the values: for example, in cell K4, use = J4 I4.
Other useful date functions are ■■ ■■ ■■
YEAR(date) MONTH(date) DAY(date)
These functions simply extract the year, month, and day from a date or cell reference that contains a date. The function TODAY( ) displays the current date. Similar to date functions, times can be formatted in a variety of ways, such as 12:26 PM, Hours:Minutes:Seconds, or in military time. The function NOW( ) displays the current time and date.
Miscellaneous Excel Functions and Tools In this section, we will illustrate a few miscellaneous Excel functions and tools that support analytics applications.
Range Names A range name is a descriptive label assigned to a cell or range of cells. Range names can help to facilitate building models on spreadsheets and understanding the formulas on which models are based. There are several ways to create range names in Excel. EXAMPLE A1.5
Using the Name Box to Create a Range Name
Suppose that we create a simple spreadsheet for computing total cost (which is the fixed cost plus unit variable cost times quantity produced) shown in Figure A1.8. We will define range names for each of the numerical cells that correspond to the labels on the left. That is, we will name cell
▲▲ Figure
A1.7
Using DATEDIF to Compute Lead Times
B3 Fixed cost, cell B4 Unit variable cost, and so on. Click on cell B3; in the Name box, type the name Fixed_cost (note the underscore), and then press Enter. Figure A1.8 shows that the name for cell B3 is displayed in the Name box. Repeat this process for each of the other numerical cells.
Appendix A1 Basic Excel Skills
▶▶ Figure
69
A1.8
Defined Range Name for Cell B3 Using the Name Box
1. Use the Name box. The name box is at the upper left of a spreadsheet between
the ribbon and the column headers. Usually, it displays the cell reference that is selected. To name a range, first select a cell or range and enter the range name in the Name box. Range names cannot contain any spaces, so it is common to use an underscore between words. 2. Use Create from Selection. This option is particularly useful when the names
you want to use are listed in the column immediately to the right or left of the cell, or in the row immediately above or below the cell or range. The next example illustrates this procedure. EXAMPL E A1.6 Using Create from Selection to Define Range Names In the Total Cost Model spreadsheet, we will use the text labels to the left of the numerical inputs as the range names. First, highlight the range A3:B5. Then, on the Formulas tab, choose Create from Selection. The box for the left column
will automatically be checked. Click OK. If you select any numerical cell, you will see the range name in the Name box as shown in Figure A1.9.
3. Use Define Name. This allows you to enter the name but also provides an option
to restrict the name to the current worksheet or allow it to be used in any worksheet of the workbook. EXAMPL E A1.7 Using Define Name to Create a Range Name In the Total Cost Model spreadsheet, select cell B3. Click Define Name on the Formulas tab. This will bring up a
▶▶ Figure
A1.9
Defined Range Name for Cell B4 Using Create from Selection
dialog that allows you to enter a range name. Figure A1.10 illustrates this. Click OK.
70
Appendix A1 Basic Excel Skills
▶▶ Figure
A1.10
Defined Range Name for Cell B3 Using Define Name
After range names have been defined, we can display a summary by clicking the Name Manager in the Formulas tab. Figure A1.11 shows this. (Note: The Name Manager button in the Formulas tab is only available in Windows. On a Mac, you can click Define Name and see a list of your range names, and then modify, add, or delete them.) This allows you to easily add, edit, or delete range names. Finally, you can apply range names to the formulas in the spreadsheet. This replaces the cell references by the names, making it easy to understand the formulas. To do this, click on the dropdown arrow next to Define Name and select Apply Names. . . . In the dialog box, select all the names you wish to use and click OK. Figure A1.12 shows the result for the total cost cell, B7. We see that the original formula = B3+B4*B5 now displays the names. ▶▶ Figure
A1.11
Name Manager Dialog (Windows)
▲▲ Figure
A1.12
Using Range Names in Formulas
Appendix A1 Basic Excel Skills
71
VALUE Function It’s not unusual to download data from the Web that looks numerical but is actually expressed as text; this is often true of data from U.S. government Web sites. One way to tell is that text is usually left justified, while numbers are right justified. The function VALUE(text) can be used to convert a text string that represents a number to a number.
Paste Special If you cut and paste a cell or a range, Excel will copy any formulas that are contained within the cell or range (using the appropriate relative or absolute addressing). In some cases, you may only want to paste the numerical values, not the formulas. This can be accomplished using the Paste Special command. From the Edit group, click the down arrow on the Paste icon to display the options and choose Paste Special. Figure A1.13 shows the Excel Paste Special dialog. There are many options, but we will only illustrate three of them. First, copy the range of cells of interest and click on the cell where you want to paste the results. ■■ ■■
To paste only the values in cells (not the formulas), select Values and then click OK. To transpose data in columns to rows and vice versa, check the Transpose box and then click OK. (Make sure you won’t overwrite any existing data!)
The Operation choices allow you to transform a range of numbers by adding, subtracting, dividing, or multiplying each number in the range by a constant. We illustrate this in the following example. EXAMPL E A1.8
Currency Conversion Using Paste Special
Figure A1.14 shows sales data in column B expressed in euros. Assume that the current conversion factor to U.S. dollars is 1 euro = $1.117. To convert these into U.S. dollars, first copy the data to column C (if you don’t, the procedure will convert the data in column B to dollars, but you will no longer have the original data). Select cell C3 corresponding
▶▶ Figure
A1.13
Paste Special Dialog
to the conversion factor. Next, select the range of data in column C and open the Paste Special dialog. Select Multiply and then press OK. The original data in euros will be converted into dollars, as shown in the figure (you will need to reformat the data as dollars).
72
Appendix A1 Basic Excel Skills
▶▶ Figure
A1.14
Converting Sales Data from Euros to Dollars
Concatenation To concatenate means to join. In many applications, you might wish to take text data that are in different columns and join them together (for example, first and last names). The Excel function CONCATENATE(text1, text2, . . . , text30) can be used to join up to 30 text strings into a single string. For example, suppose that you have a database of names, with last names in column A and first names in column B. Suppose that cell A1 contains the last name Smith, and cell B1 contains the first name John. Then CONCATENATE(B1, “ “, A1) will result in the text string John Smith. Note that the cells are separated by a field of open quotes around a blank space; this provides a space between the first and last name; otherwise, the function would have produced JohnSmith. We may also perform concatenation using Excel formulas and the operator &. For example, to concatenate the text in cells B1 and A1, enter the formula = B1&A1, which would result in JohnSmith. If we use the formula = B1&“ ”&A1, then we are simply inserting a blank between the names: John Smith. Any text string can be used; for example, =“Dr. ”&B1&“ ”&A1 would result in Dr. John Smith.
Error Values Excel displays various error values if it cannot compute a formula. These error values are the following: ■■ ■■ ■■ ■■ ■■ ■■
#DIV/0!—A formula is trying to divide by zero. #N/A—“Not available,” meaning that the formula could not return a result. #NAME?—An invalid name is used in a formula. #NUM!—An invalid argument is used in a function, such as a negative number in SQRT. #REF!—A formula contains an invalid cell reference. #VALUE!—Excel cannot compute a function because of an invalid argument.
Sometimes these are not user errors, but quirks in the data. For instance, you might have missing data in a column that is used in a denominator of a formula and obtain a #DIV/0! error. You might use the Excel function IFERROR(value, value_if_error) to display a specific value if an error is present. For example, if you are computing A1/B1, then if B1 is zero or blank, the function =IFERROR(A1/B1, “ ”) will display a blank cell instead of #DIV/0!; or = IFERROR(A1/B1, “Check the data in cell B1!”) will provide better guidance.
Appendix A1 Basic Excel Skills
73
PROBLEMS AND EXERCISES APPENDIX A1 Excel Formulas and Addressing
Excel Functions
1. Develop a spreadsheet for computing the demand
6. In the Accounting Professionals database, use Excel
for any values of the input variables in the linear demand and nonlinear demand prediction models in Examples 1.7 and 1.8 in the chapter.
functions to find a. the maximum number of years of service. b. the average number of years of service. c. the number of male employees in the database. d. the number of female employees who have a CPA.
2. The Excel file Science and Engineering Jobs shows
the number of jobs in thousands in the year 2000 and projections made for 2010 from a government study. Use the Excel file to compute the projected increase from the 2000 baseline and the percentage increase for each occupational category. 3. A new graduate has taken a job with an annual sal
ary of $60,000. She expects her salary to go up by 2.5% each year for the first five years. Her starting salary is stored in cell A4 of an Excel worksheet, and the salary increase rate is stored in cell B4. Construct a table with years 1 through 5 in cells A6:A10 and her salary in cells B6:B10. Write the formula for her salary in year 2 (in cell B7) that can be copied and pasted correctly in cells B8 through B10.
7. A company forecasts its net revenue for the next three
years as $172,800, $213,580, and $293,985. Find the net present value of these cash flows, assuming a discount rate of 4.2% 8. A pharmaceutical manufacturer has projected net
profits for a new drug that is being released to the market over the next five years: Year
Net Profit
1
($300,000,000)
2
($145,000,000)
4. Example 1.2 in the chapter described a scenario for
3
$50,000,000
new product sales that can be characterized by a formula called a Gompertz curve: ct S = aebe
4
$125,000,000
5
$530,000,000
Develop a spreadsheet for calculating sales using this formula for t = 0 to 160 in increments of 10 when a = 15,000, b = 8, and c = 0.05. 5. Return on investment (ROI) is profit divided by
investment. In marketing, ROI is determined as incremental sales times gross margin minus marketing investment, all divided by marketing investment. Suppose that a company plans to spend $3 million to place search engine ads and expects $15 million in incremental sales. Its gross margin is estimated to be 45%. a. Develop a spreadsheet to compute the marketing ROI. b. Use the spreadsheet to predict how ROI will change if the incremental sales estimate is wrong (consider a range of values above and below the expected sales).
1
A fixed cost of $80,000,000 has been incurred for research and development (in year 0). Use a spreadsheet to find the net present value of these cash flows for a discount rate of 3%. 9. The worksheet Base Data in the Excel file Credit
Risk Data provides information about 425 bank customers who had applied for loans. The data include the purpose of the loan, checking and savings account balances, number of months as a customer of the bank, months employed, gender, marital status, age, housing status and number of years at current residence, job type, and creditrisk classification by the bank.1 Use the COUNTIF function to determine (1) how many customers applied for newcar, usedcar, business, education, smallappliance, and furniture loans and (2) the number of customers with checking account balances less than $500.
Based on Efraim Turban, Ranesh Sharda, Dursun Delen, and David King, Business Intelligence: A Managerial Approach, 2nd ed. (Upper Saddle River NJ: Prentice Hall, 2011).
74
Appendix A1 Basic Excel Skills
10. The Excel file Store and Regional Sales Database
12. Define range names for all the entities in the Crebo
provides sales data for computers and peripherals showing the store identification number, sales region, item number, item description, unit price, units sold, and month when the sales were made during the fourth quarter of last year.2 Modify the spreadsheet to calculate the total sales revenue for each of the eight stores as well as each of the three sales regions.
Manufacturing Model spreadsheet and apply them to the formulas in the model. 13. Define range names for all the entities in the Hotel
Overbooking Model spreadsheet and apply them to the formulas in the model.
Miscellaneous Excel Functions and Tools 11. Define range names for all the data and model enti
ties in the BreakEven Decision Model spreadsheet and apply them to the formulas in the model.
2
Based on Kenneth C. Laudon and Jane P. Laudon, Essentials of Management Information Systems, 9th ed. (Upper Saddle River, NJ: Prentice Hall, 2011).
CHAPTER
2
Database Analytics
NAN728/Shutterstock
LEARNING OBJECTIVES ■■
■■
■■ ■■
After studying this chapter, you will be able to:
Explain the difference between a data set and a database and apply Excel range names in data files. Construct Excel tables and be able to sort and filter data. Apply the Pareto principle to analyze data. Use database functions to extract records.
■■ ■■ ■■
■■
Apply logical functions in Excel formulas. Use Excel lookup functions to make database queries. Design simple Excel templates for descriptive analytics. Use PivotTables to analyze and gain insight from data.
76
Chapter 2 Database Analytics
Most organizations begin their analytics journey using descriptive analytics tools to understand past and current performance to support customer service representatives, technical support, manufacturing, and other needs. This typically involves database analytics—querying databases and “drilling down” to better understand relationships within data, developing summarized reports, sorting data, and creating charts to visualize data (discussed in the next chapter)— techniques most commonly associated with business intelligence/information systems. These techniques provide a means of translating data into meaningful information. Data are raw numbers and facts, for example, information about customers, sales, costs, revenue, and so on. When data are manipulated, summarized, or processed in some way to provide insight and understanding, information results, which allows managers to make informed decisions. In your career, you will likely need to extract or manipulate information from a database. Typical questions that you might address or be asked to determine are the following: ■■
How many units did we sell last quarter in each store?
■■
What was our total cost of materials last month?
■■
How much profit did we make last month?
■■
What percentage of customers responded that they are highly likely to recommend us?
■■
How many defective items did each factory make?
■■
How many orders were delivered on time?
■■
How does employee satisfaction differ by job classification and region or factory? You might also use database analytics in your personal life, for example:
■■
How did my investment portfolio perform this year as compared to the previous year?
■■
In what stocks or mutual funds should I invest?
■■
Which of my fantasy sports players ranked in the top 10% of performance?
■■
How did we spend our money last year as compared to our budget?
■■
How can I compare the cost of living among various cities where I am interviewing for a job?
■■
What foods should I eat to meet my health goals or nutritional requirements?
The applications are endless! All you need is data, and data are everywhere. Many sophisticated tools, such as Microsoft Access, SQL (structured query language), and OLAP (online analytical processing) software, are available to organize and query data to answer routine business questions, and to process
Chapter 2 Database Analytics
77
data to create meaningful information. However, in this book, we won’t be delving deeply into databases or database management systems but will work with individual database files or simple data sets. Because spreadsheets are convenient tools for storing and manipulating data sets and database files, we will use them for all examples and problems. Thus, in this chapter, we focus only on basic concepts and approaches that can be implemented easily using Excel. We will introduce many useful advanced functions available in Excel and illustrate how they can be used for descriptive analytics. These functions are also useful in spreadsheet modeling, which we will study later in this book.
Data Sets and Databases A data set is simply a collection of data. Marketing survey responses, a table of historical stock prices, and a collection of measurements of dimensions of a manufactured item are examples of data sets. As we embark on statistical analysis procedures in subsequent chapters, we will see many examples of empirical data—data that come from experimentation and observation. Most data that managers work with are stored in databases. A database is a collection of related files containing records on people, places, or things. Databases provide structure to data. The people, places, or things for which we store and maintain information are called entities.1 A database is usually organized in a twodimensional table, where the columns correspond to each individual element of data (called fields, or attributes), and the rows represent records of related data elements. Empirical data, such as those obtained from questionnaires and surveys, are often organized as databases, and all the database tools that we will learn can be used. The Drout EXAMPLE 2.1
A Purchase Order Database2
Figure 2.1 shows a portion of the Excel file Purchase Orders that shows a list of orders for parts used in airplane assemblies. The fields are listed in row 3 of the database and consist of the supplier name, order number, cost,
▶▶ Figure
quantity, A/P (accounts payable) terms, and so on. Each record starting in row 4 provides the data for each of these fields for one order.
2.1
A Portion of Excel File Purchase Orders 1
Kenneth C. Laudon and Jane P. Laudon, Essentials of Management Information Systems, 9th ed. (Upper Saddle River, NJ: Prentice Hall, 2011): 159. 2 Based on Laudon and Loudon, ibid.
78
Chapter 2 Database Analytics
Advertising case at the end of this chapter is one example where you can apply database tools to empirical data. Although the databases we will use in this book are “clean,” you must be aware that many encountered in practice have typos, missing values, or other errors, which must be corrected in order to use them effectively.
Using Range Names in Databases We introduced the concept of range names in spreadsheets in the Appendix to Chapter 1. Using range names for data arrays (that is, rows, columns, or rectangular matrix arrays) can greatly simplify many database calculations. The following example shows how to do this. EXAMPLE 2.2
Defining and Using Range Names for a Database
Figure 2.2 shows a simple database for monthly sales of five products for one year (Excel file Monthly Product Sales). We will define names for each row (months) and column (product name) of data. Any of the techniques described in the Appendix to Chapter 1 can be used (we suggest that you review this material first). For instance, select the range B4:B15, which consists of the sales of Product A. In the Name box, enter Product_A and press Enter. Alternatively,
you can use Define Name to do this. Finally, you can select the entire range of data including the product names and use the Create from Selection option to define range names for all products at the same time. You may use range names to simplify calculations. For example, to find the total sales, you can use the function =SUM(Product_A) instead of = SUM(B4:B15).
Excel also provides a convenient way to include all data in a row or column in a function. For example, the range for column B is expressed as B:B; the range for row 4 is expressed as 4:4, and so on. Therefore, = SUM(B:B) will calculate the sum of all data in column B; in Figure 2.2, this will be the total sales of Product A. The function = SUM(4:4) will calculate the sum of all data in row 4; in Figure 2.2, this will be the sum of January sales for all products. Using these ranges is particularly helpful if you add new data to a spreadsheet; you would not have to adjust any defined range names. CHECK YOUR UNDERSTANDING 1. Explain the difference between a data set and a database. 2. For the Monthly Product Sales Excel file, use the Create from Selection option to define range names for all products (column data) and all months (row data). 3. For the Monthly Product Sales Excel file, use the function SUM(B:B) to calculate the total sales for Product A. What happens if you enter this function in column B? ▶▶ Figure
2.2
Monthly Product Sales Data
Chapter 2 Database Analytics
79
ANALYTICS IN PRACTICE: Using Big Data to Monitor Water Usage in Cary, North Carolina3 When the Town of Cary installed wireless meters for 60,000 customers in 2010, it knew the new technology wouldn’t just save money by eliminating manual monthly readings; the town also realized it would get more accurate and timely information about water consumption. The Aquastar wireless system reads meters once an hour—that’s 8,760 data points per customer each year instead of 12 monthly readings. The data had tremendous potential, if it could be easily consumed. The challenge was to analyze halfa billion data points on water usage and make them available, visually, to all customers, and gain a bigpicture view of water consumption to better plan future water plant expansions and promote targeted conservation efforts. Cary used advanced analytics software from SAS Institute. The ability to visually look at data by household or commercial customer, by hour, has led to some very practical applications:
■■
■■
■■
The town can notify customers of potential leaks within days. Customers can set alerts that notify them within hours if there is a spike in water usage. Customers can track their water usage online, helping them to be more proactive in conserving water.
The town estimates that by just removing the need for manual readings, the Aquastar system will save more than $10 million after the cost of the project. But the analytics component could provide even bigger savings. Already, both the town and individual citizens have saved money by catching water leaks early. As the Town of Cary continues to plan its future infrastructure needs, having accurate information on water usage will help it invest in the right amount of infrastructure at the right time. Additionally, understanding water usage will help the town if it experiences something detrimental like a drought.
Data Queries: Tables, Sorting, and Filtering
▶▶ Figure
2.3
Managers make numerous queries about data. For example, Figure 2.3 shows a portion of the Credit Risk Data Excel file, which summarizes loan application data and credit risk assessment. This database contains the purpose of the loan, the balance of checking and savings accounts, how long the applicant has been a customer and length of employment, other demographic information, and an assessment of credit risk. A manager might be interested in comparing the amount of financial assets as related to the number of months employed, finding all loans of a certain type, or the proportion of low credit risk individuals who own a home. To address these queries, we need to sort the data in some way. Extracting a set of records having certain characteristics is called filtering the data. Excel provides a convenient way of formatting databases to facilitate such types of analyses, called Tables. An Excel table allows you to use table references to perform basic calculations, as the next example illustrates.
Portion of Excel File Credit Risk Data (Based on Kenneth C. Laudon and Jane P. Laudon, Essentials of Management Information Systems, 9th ed. Upper Saddle River, NJ: Prentice Hall, 2011).
3
“Municipality puts wireless water meterreading data to work,” Copyright 2016 SAS Institute Inc. Carey, NC USA. https://www.sas.com/en_us/customers/townofcaryaquastar.html
80
Chapter 2 Database Analytics
▶▶ Figure
2.4
Portion of Credit Risk Data Formatted as an Excel Table
EXAMPLE 2.3
Creating an Excel Table
We will use the Credit Risk Data file to illustrate an Excel table. First, select the range of the data, including headers (a useful shortcut is to select the first cell in the upper left corner, then click Ctrl+Shift+down arrow, and then Ctrl+Shift+right arrow; on a Mac, use the Command key instead of Ctrl). Next, click Table from the Tables group on the Insert tab and make sure that the box for My Table Has Headers is checked. (You may also just select a cell within the table and then click on Table from the Insert menu. Excel will choose the table range for you to verify.) The table range will now be formatted and will continue automatically when new data are entered. Figure 2.4 shows a portion of
EXAMPLE 2.4
the result. Note that the rows are shaded and that each column header has a dropdown arrow to filter the data (we’ll discuss this shortly). If you click within a table, the Table Tools Design tab will appear in the ribbon, allowing you to do a variety of things, such as change the color scheme, remove duplicates, and change the formatting. A useful option in the Design tab is to add a total row. In the Table Style Options group, check the box for Total Row, and Excel will add a new row at the bottom of the table. If you click on any cell in the total row, you can select the type of calculation you want from the dropdown box; this includes SUM, AVERAGE, COUNT, MAX, and MIN.
TableBased Calculations
Suppose that in the Credit Risk Data table, we wish to calculate the total amount of savings in column C. We could, of course, simply use the function =SUM(C4:C428). However, with a table, we could use the formula =SUM(Table1[Savings]). The table name, Table1, can be found (and changed) in the Properties group of the Table Tools Design tab in Windows or in the Table tab on a
Mac. Note that Savings is the name of the header in column C. One of the advantages of doing this is that if we add new records to the table, the calculation will be updated automatically, and we don’t have to change the range in the formula or get a wrong result if we forget to. As another example, we could find the number of home owners using the function =COUNTIF(Table1[Housing], “Own”).
If you add additional records at the end of the table, they will automatically be included and formatted, and if you create a chart based on the data, the chart will automatically be updated if you add new records.
Sorting Data in Excel Excel provides many ways to sort lists by rows or column or in ascending or descending order and using custom sorting schemes. The sort buttons in Excel can be found under the Data tab (see Figure 2.5; the Mac Data tab ribbon is similar). Select a single cell in the column you want to sort on and click the “AZ down arrow” button to sort from smallest to largest or the “AZ up arrow” button to sort from largest to smallest. You may also click the Sort button to specify criteria for more advanced sorting capabilities.
Chapter 2 Database Analytics
▶▶ Figure
81
2.5
Excel Data Tab Ribbon
EXAMPLE 2.5
Sorting Data in the Purchase Orders Database
In the Purchase Orders database, suppose we wish to sort the data by supplier. Click on any cell in column A of the data (but not the header cell A3) and then the “AZ down” button in the Data tab. Excel will select the entire range of
the data and sort by name of supplier in column A, a portion of which is shown in Figure 2.6. This allows you to easily identify the records that correspond to all orders from a particular supplier.
Pareto Analysis Pareto analysis is a term named after an Italian economist, Vilfredo Pareto, who, in 1906, observed that a large proportion of the wealth in Italy was owned by a relatively small proportion of the people. The Pareto principle is often seen in many business situations. For example, a large percentage of sales usually comes from a small percentage of customers, a large percentage of quality defects stems from just a couple of sources, or a large percentage of inventory value corresponds to a small percentage of items. As a result, the Pareto principle is also often called the “80–20 rule,” referring to the generic situation in which 80% of some output comes from 20% of some input. A Pareto analysis relies on sorting data and calculating the cumulative percentage of the characteristic of interest. EXAMPLE 2.6
Applying the Pareto Principle
The Excel file Bicycle Inventory lists the inventory of bicycle models in a sporting goods store.4 To conduct a Pareto analysis, we first compute the inventory value of each product by multiplying the quantity on hand by the purchase cost; this is the amount invested in the items that are currently in stock. Then we sort the data in decreasing order of inventory value and compute the percentage of the total
▶▶ Figure
inventory value for each product and the cumulative percentage. See columns G through I in Figure 2.7. We see that about 75% of the inventory value is accounted for by less than 40% (9 of 24) of the items. If these highvalue inventories aren’t selling well, the store manager may wish to keep fewer in stock.
2.6
Portion of Purchase Orders Database Sorted by Supplier Name 4
Based on Kenneth C. Laudon and Jane P. Laudon, Essentials of Management Information Systems, 9th ed. (Upper Saddle River, NJ: Prentice Hall, 2011).
82
Chapter 2 Database Analytics
▶▶ Figure
2.7
Pareto Analysis of Bicycle Inventory
Filtering Data For large data files, finding a particular subset of records that meet certain characteristics by sorting can be tedious. Filtering simplifies this process. Excel provides two filtering tools: AutoFilter for simple criteria and Advanced Filter for more complex criteria. These tools are best understood by working through some examples. EXAMPLE 2.7
Filtering Records by Item Description
In the Purchase Orders database, suppose we are interested in extracting all records corresponding to the item boltnut package. First, select any cell within the database. Then, from the Excel Data tab, click on Filter. A dropdown arrow will then be displayed on the right side of each header column. Clicking on one of these will display a dropdown box. These are the options for filtering on that ▶▶ Figure
2.8
Selecting Records for BoltNut Package
column of data. Click the one next to the Item Description header. Uncheck the box for Select All and then check the box corresponding to the boltnut package, as shown in Figure 2.8. Click the OK button, and the Filter tool will display only those orders for this item (Figure 2.9). To restore the original data file, click on the dropdown arrow again and then click Clear filter from “Item Description.”
Chapter 2 Database Analytics
▶▶ Figure
83
2.9
Filter Results for BoltNut Package
EXAMPLE 2.8
Filtering Records by Item Cost
In this example, suppose we wish to identify all records in the Purchase Orders database whose item cost is at least $200. First, click on the dropdown arrow in the Item Cost column and position the cursor over Numbers Filter. This displays a list of options, as shown in Figure 2.10. Select Greater Than Or Equal To . . . from the list. (The Mac
interface is slightly different, but has the same capabilities.) This brings up a Custom AutoFilter dialog (Figure 2.11) that allows you to specify up to two specific criteria using “and” and “or” logic. Enter 200 in the box as shown in Figure 2.11 and then click OK. The tool will display all records having an item cost of $200 or more.
AutoFilter creates filtering criteria based on the type of data being filtered. For instance, in Figure 2.10 we see that the Number Filters menu list includes numerical criteria such as “equals,” “does not equal,” and so on. If you choose to filter on Order Date or Arrival Date, the AutoFilter tools will display a different Date Filters menu list for filtering that includes “tomorrow,” “next week,” “year to date,” and so on. AutoFilter can be used sequentially to “drill down” into the data. For example, after filtering the results by boltnut package in Figure 2.9, we could then filter by order date and select all orders processed in September. ▶▶ Figure
2.10
Selecting Records for Item Cost Filtering
84
Chapter 2 Database Analytics
▶▶ Figure
2.11
Custom AutoFilter Dialog
Caution! The Filter tool does not extract the records; it simply hides the records that don’t match the criteria. This means that if you highlight a range of filtered data to compute a sum or average, for example, you will get the results for all records, including those that are hidden from view. However, you can copy and paste the filtered data to another range or Excel worksheet, which then uses only the filtered records, and get the correct results. The Advanced Filter provides a way of explicitly defining criteria by which to filter a database. To do this, first copy the headers from the database to an open location in the worksheet. Under the headers, specify the criteria that you want to use to filter the data. Multiple criteria in the same row are logically joined by “and,” while criteria in rows are joined by “or.” In general, criteria for numerical values can include =, 7 , 6 , 6 7 (not equal), 7 =, or 6 = . EXAMPLE 2.9
Using the Advanced Filter
We will use the Purchase Orders database. Figure 2.12 shows the criteria by which we want to filter the records. These criteria will find all the records for Hulkey Fasteners having order quantities that exceed 5,000 and order dates before 9/1/11, as well as all records for Steelpin Inc. with order quantities less than 5,000 and order dates before 9/1/11. You need not include columns that have no criteria; in this case, we could have simply listed the columns
Supplier, Quantity, and Order Date along with the specified criteria. To use the Advanced Filter, choose Advanced next to the Filter in the Data tab. In the dialog box, enter the list range of the database and the criteria range (see Figure 2.13). Figure 2.14 shows the results; note that only the records meeting the criteria are displayed.
Database Functions You are already familiar with basic Excel functions such as SUM, AVERAGE, COUNT, and so on. Database functions start with a “D” (for example, DSUM, DAVERAGE, DCOUNT) and allow you to specify criteria that limit the calculations to a subset of records
▶▶ Figure
2.12
Criteria for Using the Advanced Filter for the Purchase Orders Database
Chapter 2 Database Analytics
▶▶ Figure
85
2.13
Advanced Filter Dialog
▶▶ Figure
2.14
Advanced Filter Results for the Purchase Orders Database
in a database using the same format as the Advanced Filter. You can find all database functions by selecting the category Database in the Insert Function dialog. For example, the syntax for the DSUM function is DSUM(database, field, criteria). Database is the range that includes the column labels; field is the column name that contains the values to sum, enclosed in quotation marks, or a reference to the column name; and criteria is the range that specifies the records you want to sum (the same format as in Figure 2.12).
EXAMPLE 2.10
Using a Database Function
In the Purchase Orders database, suppose that we wish to find the total cost of all orders that meet the criteria specified in Figure 2.12. Figure 2.15 shows the application of the DSUM function. In this function, Criteria is a range name that corresponds to the criteria range A3:J5, in the
worksheet. Note that the function references cell G3 for the field, which corresponds to cost per order (we might also have created a range name for the entire database). Alternatively, we could have used the function =DSUM(A8:J102, “Cost per order”, A3:J5).
86
Chapter 2 Database Analytics
▶▶ Figure
2.15
Using the DSUM Function
CHECK YOUR UNDERSTANDING 1. What does the term filtering mean in the context of a database? 2. Convert the Credit Risk database to an Excel table, and use a tablebased function to count the number of records designated as having high credit risk. 3. Sort the data in the Purchase Orders database from lowest cost per order to highest. 4. Use both the AutoFilter and the Advanced Filter to identify all orders for boltnut packages that are processed in September in the Purchase Orders database.
ANALYTICS IN PRACTICE: Discovering the Value of Database Allders International specializes in dutyfree operations with 82 taxfree retail outlets throughout Europe, including shops in airports and seaports and on crosschannel ferries. Like most retail outlets, Allders International must track masses of pointofsale data to assist in inventory and productmix decisions. Which items to stock at each of its outlets can have a significant impact on the firm’s profitability. To assist them, they implemented a computerbased data warehouse to maintain the data. Prior to doing this, they had to analyze large quantities of paperbased data. This manual process was so overwhelming and timeconsuming that the analyses were often too late to provide useful information for their decisions. The data warehouse allowed the company to make simple queries, such as finding the performance of a particular item across all retail outlets or the financial performance of a particular outlet, quickly and easily. This allowed them to identify which inventory items or outlets were underperforming. For instance, a Pareto analysis of its product lines (groups of similar items)
5
2jenn/Shutterstock
Analytics at Allders International5
found that about 20% of the product lines were generating 80% of the profits. This allowed them to selectively eliminate some of the items from the other 80% of the product lines, which freed up shelf space for more profitable items and reduced inventory and supplier costs.
Based on Stephen Pass, “Discovering Value in a Mountain of Data,” OR/MS Today, 24, 5 (December 1997): 24–28. (OR/MS Today was the predecessor of Analytics magazine.)
Chapter 2 Database Analytics
87
Logical Functions Logical functions depend on whether one or more conditions are true or false. A condition is a statement about the value of a cell, either numeric or text. Three useful logical functions in business analytics applications are the following: IF(condition, value if true, value if false)—a logical function that returns one value if the condition is true and another if the condition is false AND(condition 1, condition 2, . . . )—a logical function that returns TRUE if all conditions are true and FALSE if not OR(condition 1, condition 2, . . . )—a logical function that returns TRUE if any condition is true and FALSE if not The IF function, IF(condition, value if true, value if false), allows you to choose one of two values to enter into a cell. If the specified condition is true, value if true will be put in the cell. If the condition is false, value if false will be entered. Value if true and value if false can be a number or a text string enclosed in quotes. For example, if cell C2 contains the function =IF(A8 = 2,7,12), it states that if the value in cell A8 is 2, the number 7 will be assigned to cell C2; if the value in cell A8 is not 2, the number 12 will be assigned to cell C2. Conditions may include the following: = 7 6 7= 6= 67
equal to greater than less than greater than or equal to less than or equal to not equal to
Note that if a blank is used between quotes, “ ”, then the result will simply be a blank cell. This is often useful to create a clean spreadsheet. AND and OR functions simply return the values of true or false if all or at least one of multiple conditions are met, respectively. You may use AND and OR functions as the condition within an IF function; for example, =IF(AND(B1=3,C1=5),12,22). Here, if cell B1=3 and cell C1=5, then the value of the function is 12; otherwise, it is 22.
EXAMPLE 2.11
Using the IF Function
In the Purchase Orders database, suppose that the aircraft component manufacturer considers any order of 10,000 units or more to be large, whereas any other order size is considered to be small. We may use the IF function to classify the orders. First, create a new column in the spreadsheet for the order size, say, column K. In cell K4, use the formula = IF(F4 7 = 10000, “Large”, “Small”) This function will return the value Large in cell K4 if the order size in cell F4 is 10,000 or more; otherwise, it returns
the value Small. Further, suppose that large orders with a total cost of at least $25,000 are considered critical. We may flag these orders as critical by using the function in cell L4: = IF(AND(K4 = “Large”, G4 7 = 25000), “Critical”, “ ”) Note that we use open quotes to return a blank cell if the order is not critical. After copying these formulas down the columns, Figure 2.16 shows a portion of the results.
88
Chapter 2 Database Analytics
▶▶ Figure
2.16
Classifying Order Sizes Using the IF Function
You may “nest” up to seven IF functions by replacing value if true or value if false in an IF function with another IF function: =IF(A8=2,(IF(B3=5, “YES”, “”)),15) This says that if cell A8 equals 2, then check the contents of cell B3. If cell B3 is 5, then the value of the function is the text string YES; if not, it is a blank space. However, if cell A8 is not 2, then the value of the function is 15 no matter what cell B3 is. One tip to do this easily is to write out each IF formula first, then successively embed them into one formula.
EXAMPLE 2.12
Calculating the Price of Quantity Discounts
Suppose that a company offers quantity discounts on purchases. For quantities of 1,000 or less, the unit price is $10; for quantities of 1,001 to 5,000, the unit price is $9.00; and for quantities that exceed 5,000, the unit price is $7.50. We may use nested IF functions to compute the total cost for any purchase quantity. Here is the logic. Let Q represent the purchase quantity. If Q 6 = 1,000, the price is $10 * Q; if this is not true, then we need to check if Q 6 = 5,000. If so, the price will be $9.00 * Q. Finally, if this is false, then the quantity must exceed 5,000, and the price will be $7.50 * Q. Write the first IF statement as IF1Q 6 = 1,000, Q*10, value if false2
▶▶ Figure
2.17
Spreadsheet Implementation of Quantity Discount Calculation
If this is false, we would have the following IF statement: IF1Q 6 = 5,000, Q*9, value if false2 If this is false, the value must be Q times 7.5. Now substitute: IF(Q 6 = 1,000, Q*10, value if false) IF(Q 6 = 1,000, Q*10, IF(Q6 = 5,000, Q*9, value if false)) IF(Q 6 = 1,000, Q*10, IF(Q6 = 5,000, Q*9, Q*7.5))
Figure 2.17 shows the formula used in the spreadsheet implementation.
Chapter 2 Database Analytics
89
CHECK YOUR UNDERSTANDING 1. Write an IF function that returns the word “Valid” if the value of cell B16 is greater than or equal to 0 and the word “Invalid” if not. 2. Explain the difference between the AND and OR functions. 3. Write an IF function that performs the following and explain how it should be implemented: If the contents of cell D10 is the text string “Invoice received” and the contents of cell E10 is “Paid,” then set the contents of cell F10 to “Order Completed”; otherwise, set the contents of cell F10 to “Order Open.”
Lookup Functions for Database Queries Excel provides some useful lookup functions for finding specific data in a spreadsheet. These are the following: VLOOKUP(lookup_value, table_array, col_index_num, [range lookup]) looks up a value in the leftmost column of a table (specified by the table_array) and returns a value in the same row from a column you specify (col_index_num). HLOOKUP(lookup_value, table_array, row_index_num, [range lookup]) looks up a value in the top row of a table and returns a value in the same column from a row you specify. INDEX(array, row_num, col_num) returns a value or reference of the cell at the intersection of a particular row and column in a given range. MATCH(lookup_value, lookup_array, match_type) returns the relative position of an item in an array that matches a specified value in a specified order. CHOOSE(index_num, value1, value2, . . . ) returns a value from a list based on the position in the list, specified by index_num. In the VLOOKUP and HLOOKUP functions, range lookup is optional. If this is omitted or set as True, then the first column of the table must be sorted in ascending numerical EXAMPLE 2.13
Using the VLOOKUP Function
A database of sales transactions for a firm that sells instructional fitness books and DVDs is provided in the Excel file Sales Transactions. The database is sorted by customer ID, and a portion of it is shown in Figure 2.18. Suppose that a customer calls a representative about a payment issue. The representative finds the customer ID—for example, 10007—and needs to look up the type of payment and transaction code. We may use the VLOOKUP function to do this. In the function =VLOOKUP(lookup_value, table_array, col_index_num), lookup_value represents the customer ID. The table_array is the range of the data in the spreadsheet; in this case, it is the range A4:H475. The value for col_index_num represents the column in the table range we wish to retrieve. For the type of payment, this is column 3; for the transaction code, this is column 4. Note that the first column is already sorted in ascending numerical order, so we can either omit the range lookup argument or set it as
True. Thus, if we enter the formula below in any blank cell of the spreadsheet (we recommend using absolute references for arrays) = VLOOKUP(10007,$A$4:$H$475,3) it returns the payment type Credit. If we use the following formula: = VLOOKUP(10007,$A$4:$H$475,4) the function returns the transaction code 80103311. Now suppose the database was sorted by transaction code so that the customer ID column is no longer in ascending numerical order, as shown in Figure 2.19. If we use the function =VLOOKUP (10007,$A$4:$H$475,4, True), Excel returns #N/A. However, if we change the range lookup argument to False, then the function returns the correct value of the transaction code.
90
Chapter 2 Database Analytics
▶▶ Figure
2.18
Portion of Sales Transactions Data Sorted by Customer ID
▶▶ Figure
2.19
Portion of Sales Transactions Data Sorted by Transaction Code
order. If an exact match for the lookup_value is found in the first column, then Excel will return the value of the col_index_num of that row. If an exact match is not found, Excel will choose the row with the largest value in the first column that is less than the lookup_ value. If range lookup is False, then Excel seeks an exact match in the first column of the table range. If no exact match is found, Excel will return #N/A (not available). We recommend that you specify the range lookup to avoid errors. The HLOOKUP function works in a similar fashion. For most spreadsheet databases, we would normally need to use the VLOOKUP function. In some modeling situations, however, the HLOOKUP function can be useful if the data are arranged column by column rather than row by row. The INDEX function works as a lookup procedure by returning the value in a particular row and column of an array. For example, in the Sales Transactions database, =INDEX($A$4:$H$475, 7, 4) would retrieve the transaction code 80103311 that is in the 7th row and 4th column of the data array (see Figure 2.18), as the VLOOKUP function did in Example 2.13. The difference is that it relies on the row number rather than the actual value of the customer ID. In the MATCH function, lookup_value is the value that you want to match in lookup_ array, which is the range of cells being searched. The match_type is either 1, 0, or 1. The default is 1. If match type = 1, then the function finds the largest value that is less than or equal to lookup_value; however, the values in the lookup_array must be placed in ascending order. If match type = 0, MATCH finds the first value that is exactly equal to lookup_ value; in this case the values in the lookup_array can be in any order. If match type = 1, then the function finds the smallest value that is greater than or equal to lookup_value; in this case the values in the lookup_array must be placed in descending order. Example 2.14 shows how the INDEX and MATCH functions can be used. The VLOOKUP function will not work if you want to look up something to the left of a specified range (because it uses the first column of the range to find the lookup value). However, we can use the INDEX and MATCH functions easily to do this, as Example 2.14 shows.
Chapter 2 Database Analytics
EXAMPLE 2.14
91
Using INDEX and MATCH Functions for Database Queries
Figure 2.20 shows the data in the Excel file Monthly Product Sales Queries. Suppose we wish to design a simple query application to input the month and product name, and retrieve the corresponding sales. The three additional worksheets in the workbook show how to do this in three different ways. The Query1 worksheet (see Figure 2.21) uses the VLOOKUP function with embedded IF statements. The formula in cell I8 is
In this case, the MATCH function is used to identify the column in the table corresponding to the product name in cell I6. Note the use of the “ +1” to shift the relative column number of the product to the correct column number in the lookup table. Finally, the Query3 worksheet (also not shown here) uses only INDEX and MATCH functions in cell I8. The formula in cell I8 is
= VLOOKUP(I5, A4:F15, IF(I6 =“A”, 2, IF(I6 = “B”, 3, IF(I6 = “C”,4, IF(I6 = “D”, 5, IF(I6 = “E”, 6))))),FALSE)
=INDEX(A4:F15,MATCH(I5,A4:A15,0),MATCH(I6,A3:F3,0))
The IF functions are used to determine the column in the lookup table to use and, as you can see, is somewhat complex, especially if the table were much larger. The Query2 worksheet (not shown here; see the Excel workbook) uses the VLOOKUP and MATCH functions in cell I8. The formula in cell I8 is =VLOOKUP(I5,A4:F15,MATCH(I6,B3:F3,0) + 1,FALSE)
▶▶ Figure
2.20
Monthly Product Sales Queries Workbook
▶▶ Figure
2.21
Query1 Worksheet in Monthly Product Sales Queries Workbook
The MATCH functions are used as arguments in the INDEX function to identify the row and column numbers in the table based on the month and product name. The INDEX function then retrieves the value in the corresponding row and column. This is perhaps the cleanest formula of the three. By studying these examples carefully, you will better understand how to use these functions in other applications.
92
Chapter 2 Database Analytics
EXAMPLE 2.15
Using INDEX and MATCH for a Left Table Lookup
Suppose that, in the Sales Transactions database, we wish to find the customer ID associated with a specific transaction code. Refer back to Figure 2.18 or the Excel workbook. Suppose that we enter the transaction code in cell K2 and want to display the customer ID in cell K4. Use the following formula in cell K4:
=INDEX(A4:A475,MATCH(K2,D4:D475,0),1) Here, the MATCH function is used to identify the row number in the table range that matches the transaction code exactly, and the INDEX function uses this row number and column 1 to identify the associated customer ID.
To illustrate the CHOOSE function, suppose that in the monthly product sales data (see Figure 2.20), we want to select the January sales of product B. We could use =CHOOSE(2,B4,C4,D4,E4,F4), which would return the value 5,554. We may also specify a list of text names, such as =CHOOSE(2,“A”,“B”,“C”,“D”,“E”), which would return B. CHECK YOUR UNDERSTANDING 1. Explain the purpose of the VLOOKUP, INDEX, and MATCH functions. 2. In the Purchase Orders database, what will =VLOOKUP(4111,C4:J97, 5, TRUE) and =VLOOKUP(4111,C4:J97, 5, FALSE) find? Explain the difference. 3. Write an Excel function that is entered in cell J2 that will find the amount of an order for any customer ID in the Sales Transactions database using (1) only VLOOKUP, and (2) only the INDEX and MATCH functions.
Excel Template Design Many database queries are repetitive; think of customer service representatives who must look up order information, prices, and so on. We may use the logical and lookup functions that we studied in the previous section to create userfriendly Excel templates for repetitive database queries. We did this in Example 2.14 for finding sales for a specific month and product name. In this section, we discuss some design approaches and Excel tools. An Excel template would typically have input cells, result cells, and possibly intermediate calculations. Templates should be “clean,” well organized, and easy to use. Validation tools should be used to ensure that users do not inadvertently make errors in input data. Visualizations, such as charts, which will be discussed in the next chapter, can be included for additional insight. EXAMPLE 2.16
A Tax Bracket Calculator
The table below shows the 2016 U.S. federal income tax rates for the four different types of filing status.
2. Married filing jointly or qualifying widow/widower
3. Married filing seperately
4. Head of household Tax rate
Up to $9,275
Up to $18,550
Up to $9,275
Up to $13,250
10%
$9,276 to $37,650
$18,551 to $75,300
$9,276 to $37,650
$13,251 to $50,400
15%
$37,651 to $91,150
$75,301 to $151,900
$37,651 to $75,950
$50,401 to $130,150
25%
$91,151 to $190,150
$151,901 to $231,450
$75,951 to $115,725
$130,151 to $210,800
28%
$190,151 to $413,350
$231,451 to $413,350
$115,726 to $206,675
$210,801 to $413,350
33%
$413,351 to $415,050
$413,351 to $466,950
$206,676 to $233,475
$413,351 to $441,000
35%
$415,051 or more
$466,951 or more
$233,476 or more
$441,001 or more
39.60%
1. Single
Chapter 2 Database Analytics
We will create a simple template (see the Excel file Tax Bracket Template) that allows an individual to enter their filing status and taxable income and then returns the corresponding tax bracket. This is shown in Figure 2.22. We can use a lookup table to identify the tax rate. First, note that the main worksheet does not include the lookup table to keep the template clean. The input and results cells are colorcoded to separate inputs from the result. We used a nested IF statement
93
to select the proper lookup range based on the filing status code. An alternative way of doing this would be to use the CHOOSE function as follows (“Lookup” is the name of the worksheet containing the lookup table): =CHOOSE($B$9,VLOOKUP($B$10,Lookup!$A$2:$E$8,5), VLOOKUP($B$10,Lookup!$B$2:$E$8,4),VLOOKUP($B$10, Lookup!$C$2:$E$8,3),VLOOKUP($B$10,Lookup!$D$2:$E$8,2))
Data Validation Tools Excel provides various data validation tools to reduce the chances that users will make a mistake. First, select the cell range for which data validation will be applied, and then choose Data Validation from the Data Tools group on the Data tab in Excel 2016, or select Validation from the Data menu on the a Mac.
▶▶ Figure
2.22
Excel Template for Identifying Tax Bracket
94
Chapter 2 Database Analytics
▶▶ Figure
2.23
Data Validation Dialog
▶▶ Figure
2.24
Data Validation Error Alert
EXAMPLE 2.17
Applying Data Validation
For the Tax Bracket template, choose cell B9 (filing status). In the Data Validation dialog, specify that the value must be a whole number between 1 and 4, as shown in Figure 2.23. If the input value does not meet these criteria, a default error alert will pop up (see Figure 2.24). This
can be customized from the Error Alert tab in the Data Validation dialog. You may also use the Input Message tab to create a prompt to guide the user when an input cell is selected.
Another way of applying data validation for numerical inputs is to use the ISNUMBER function. This function returns TRUE if the cell or range contains a number; otherwise, it returns FALSE. This can help to avoid any typing mistakes resulting from inadvertently pressing a letter or symbol key.
EXAMPLE 2.18
Using the ISNUMBER Function for Data Validation
In the Tax template, we will use ISNUMBER to ensure that the taxable income in cell B10 is numeric. In the Data Validation dialog, choose Custom, and in the Formula
box, enter the formula =ISNUMBER(B10), as shown in Figure 2.25. An error alert message will pop up if an invalid entry is made.
▶▶ Figure
Chapter 2 Database Analytics
95
2.25
ISNUMBER Data Validation
Form Controls Form controls are buttons, boxes, and other mechanisms for inputting or changing data on spreadsheets easily that can be used to design userfriendly spreadsheets. They allow the user to more easily interface with models to enter or change data without the potential of inadvertently introducing errors in formulas. To use form controls, you must first activate the Developer tab on the ribbon. Click the File tab, then Options, and then Customize Ribbon. Under Customize the Ribbon, make sure that Main Tabs is displayed in the dropdown box, and then click the check box next to Developer (which is typically unchecked in a standard Excel installation). You will see the new tab in the Excel ribbon, as shown in Figure 2.26. If you click the Insert button in the Controls group, you will see the form controls available (do not confuse these with the Active X Controls in the same menu!). On a Mac, select Excel 7 Preferences 7 View and check the box for Developer tab. The Mac ribbon is different and shows graphics of the form controls. The most common form controls are ■■ ■■ ■■ ■■ ■■ ■■ ■■
▶▶ Figure
2.26
Spin button—a button used to increase or decrease a numerical value Scroll bar—a slider used to change a numerical value Check box—a box used to select or deselect a scenario Option button—a radio button used to select an option List box—a box that provides a list of options Combo box—a box that provides an expandable list of options Group box—a box that can hold a group of controls
Windows
Excel 2016 Developer Tab
Mac
96
Chapter 2 Database Analytics
▶▶ Figure
2.27
Tax Bracket Template with Form Controls
To insert a form control, click the Insert button in the Controls tab under the Developer menu, click on the control you want to use, and then click within your worksheet. The following example shows how to use both a spin button and scroll bar in the Tax Bracket template. EXAMPLE 2.19
Using Form Controls in the Tax Bracket Template
In the Tax Bracket template, we will use a spin button for the filing status and a scroll bar for the taxable income (in unit increments of $100, from $0 through $500,000). The completed spreadsheet is shown in Figure 2.27. First, click the Insert button in the Controls group of the Developer tab, select the spin button, click it, and then click somewhere in the worksheet. The spin button (and any form control) can be resized by dragging the handles along the edge and moving the button within the worksheet. Move it to a convenient location, and enter the name you wish to use (such as Filing Status) adjacent to it. Next, right click the spin button and select Format Control. You will see the dialog box shown in Figure 2.28. Enter the values shown and click OK. Now if you click the up or down buttons, the value in cell B9 will change within the specified range.
Next, repeat this process by inserting the scroll bar next to the taxable income cell. Right click the scroll bar and select Format Control. For a scroll bar, the maximum value is limited to 30,000, so we need to scale the parameters so that the taxable income can vary up to $500,000. Set the parameters in the dialog for a minimum value of 0 and maximum value of 5,000, with an incremental change of 1. Choose a blank cell for the cell link (say, D9). Then, in cell B10, enter the formula =D9*100. Now when the scroll bar is moved, the income will change from 0 to $500,000. You may also click the arrows on the end of the scroll bar to finetune the value. Using these controls, you can easily change the inputs in the template.
Form controls only allow integer increments, so you would also have to scale values to change a number by a fractional value. For example, suppose that we want to use a spin button to change an interest rate that is in cell B8 from 0% to 10% in increments of 0.1% (that is, 0.001). Choose some empty cell, say, C8, and enter a value between 0 and 100 in it. Then enter the formula = C8/1000 in cell B8. Note that if the value in C8 = 40, for example, then the value in cell B8 will be 40/1000 = 0.04, or 4%. Then as the value in cell C8 changes by 1, the value in cell B8 changes by 1/1,000, or 0.1%. In the Format Control dialog, specify the minimum value at 0 and the maximum value at 100 and link the button to cell C8. Now as you click the up or down arrows on the spin button, the value in cell C8 changes by 1 and the value in cell B8 changes by 0.1%.
Chapter 2 Database Analytics
▶▶ Figure
97
2.28
Format Control Dialog
Other form controls can also be used; for instance, in the Tax Bracket template, you might use a list box to select the filing status. We encourage you to experiment and identify creative ways to use them. The next example shows the use of the combo box. EXAMPLE 2.20
Using a Combo Box
Figure 2.29 shows the application of a combo box to compute the total sales for any month in the Monthly Product Sales database. We set the cell link to H1; this provides the row number of the month in the data array to use in an INDEX function to compute total sales. The formula in cell H7 is =SUM(INDEX(B4:F15,H1,0)). Using a 0 as the
▶▶ FIGURE
2.29
Combo Box Application
column number references the entire row, so the function sums all the columns in the row listed in cell H1. In the Format Control dialog, set the Drop Down lines to 12; this will display all 12 months when the arrow button is clicked. A smaller number will require you to scroll down to seem them all.
98
Chapter 2 Database Analytics
CHECK YOUR UNDERSTANDING 1. State the purpose of an Excel template and the key properties that it should have. 2. What is the purpose of Data Validation in Excel? 3. What will the function =ISNUMBER(A1) return when the contents of cell A1 is either 10 or the word ten? 4. Explain the value of form controls in Excel templates.
PivotTables Excel provides a powerful tool for distilling a complex database into meaningful information: PivotTables (yes, it is one word!). PivotTables allows you to create custom summaries and charts of key information in the data and to drill down into a large set of data in numerous ways. In this section, we will introduce PivotTables in the context of databases; however, we will also use them in other chapters to facilitate data visualization and statistical analysis of data sets. To apply PivotTables, first, select any cell in the database and choose PivotTable under the Insert tab and follow the steps of the wizard. We will use the Sales Transactions database (see Figure 2.18). Excel first asks you to select a table or range of data; if you click on any cell within the database before inserting a PivotTable, Excel will default to the complete range of your data. You may either put the PivotTable into a new worksheet or in a blank range of the existing worksheet. Excel then creates a blank PivotTable, as shown in Figure 2.30. In the PivotTable Fields window on the right side of Figure 2.30 is a list of the fields that correspond to the headers in the database. You select which ones you want to include, as row labels, column labels, values, or a filter. You should first decide what types of tables you wish to create—that is, what fields you want for the rows, columns, and data values. EXAMPLE 2.21
Creating a PivotTable
We will use a PivotTable for the Sales Transactions database to find the total amount of revenue in each region, and also the total revenue in each region for each product. If you drag the field Region from the PivotTable Fields list in Figure 2.30 to the Rows area and the field Amount into the π Values area, you will create the PivotTable shown in Figure 2.31. The values in column B are the sum of the sales revenues for each region. You may select the range of the value, right click, and format the cells as currency to make the PivotTable more meaningful. You may also replace the names in the headers; for instance, change Row Labels to Region and Sum of Amount to Revenue. The dropdown arrow next to Row Labels allows you to filter the results and display only selected regions. These modifications are shown in Figure 2.32. The beauty of PivotTables is that if you wish to change the analysis, you can simply drag the field
names out of the field areas (or just uncheck the boxes in the PivotTable Field list) or drag them to different areas. For example, if you drag the Product field into the Columns area, you will create the PivotTable in Figure 2.33, showing the breakdown of revenues by product for each region. Another option is to drag Product to the Rows area, as shown in Figure 2.34. The order in which the fields are listed will also change the view (try moving Region after Product). The best way to learn about PivotTables is simply to experiment with them. Dragging a field into the Filters area in the PivotTable Field list allows you to add a third dimension to your analysis. For example, moving Payment to the Filters area allows you to filter revenues by PayPal or credit, as shown in Figure 2.35.
▶▶ Figure
Chapter 2 Database Analytics
2.30
Blank PivotTable
▶▶ Figure
2.31
PivotTable for Total Revenue by Region
▶▶ Figure
2.32
Reformatted PivotTable for Total Revenue by Region
99
100
Chapter 2 Database Analytics
▶▶ Figure
2.33
Revenue Breakdown by Region and Product
▶▶ Figure
2.34
Alternate PivotTable View of Region and Product Revenue
PivotTable Customization When you drag a field into the π Values area, the PivotTable defaults to the sum of the values in that field. Often, this is not what you want, but it can be easily changed. For
▶▶ Figure
Chapter 2 Database Analytics
101
2.35
PivotTable Filtered by Credit Payment
example, suppose that you wanted a count of the number of transactions for each region and product. In Figure 2.31, click the dropdown arrow next to Revenue in the π Values area and choose Value Field Settings. In the dialog shown in Figure 2.36, select Count (many other options exist that perform statistical calculations; we will use these in Chapter 4). This results in the PivotTable shown in Figure 2.37 (after reformatting the values to numbers instead of currency). Now we can see how many books and DVDs were sold in each region. An easier way to reformat data in a PivotTable is to select any cell, right click, and choose Value Field Settings. Notice the Number Format button at the bottom left in Figure 2.36. If you click this, you will be able to choose a format for all the data in the PivotTable. ▶▶ Figure
2.36
Value Field Settings Dialog
102
Chapter 2 Database Analytics
▶▶ Figure
2.37
PivotTable for Count of Regional Sales by Product
You may express the data in a PivotTable in various percentage views, such as the percent of the grand total, percent of row total, percent of column total, and other options. In the Value Field Settings dialog, click the tab Show Values As. Select the option in the dropdown box; Figure 2.38 illustrates the selection of % of Row Total. Figure 2.39 shows the results; we now can see the percentage breakdown of product sales for each individual region. Other options for customizing PivotTables can be found in the PivotTable Tools Design tab. Again, we invite you to experiment with them.
▶▶ Figure
2.38
Show Values Tab in Value Field Settings
Chapter 2 Database Analytics
▶▶ Figure
103
2.39
Percentage of Product Sales by Region
Slicers Excel has a tool called slicers—for drilling down to “slice” a PivotTable and display a subset of data. Slicers simplify the process of filtering a PivotTable and are very useful in presentations to quickly show alternate PivotTable views. To create a slicer for any of the columns in the database, click on the PivotTable and choose Insert Slicer from the Analyze tab in the PivotTable Tools ribbon in Windows, or from the PivotTable Analyze tab on a Mac. EXAMPLE 2.22
Using Slicers
For the PivotTable we created in Figure 2.33 for the revenues by region and product, let us insert a slicer for the source of the transaction, as shown in Figure 2.40. From the Insert Slicers dialog, check the box for Source. This results in the slicer window shown in Figure 2.40. If you click on one of the Source buttons, Email or Web, the
PivotTable reflects only those records corresponding to that source. In Figure 2.40, we are viewing only the total revenues generated through Web sales. If you click on the small red “x” by the filter in the top right of the slicer, you would clear the filter and the PivotTable would show revenues from all sources, as in Figure 2.33.
CHECK YOUR UNDERSTANDING 1. What is the purpose of a PivotTable? 2. Explain how to construct a PivotTable for the Sales Transactions database so that it shows the count of the number of transactions using PayPal and credit for each region. 3. What is a slicer? How does it simplify the use of PivotTables?
▶▶ Figure
2.40
Revenues Sliced by Web Sales
104
Chapter 2 Database Analytics
KEY TERMS Data set Data validation Database Database functions Empirical data Excel table Excel template
Filtering Form controls Logical functions Lookup functions Pareto analysis PivotTables Slicers
CHAPTER 2 TECHNOLOGY HELP Useful Excel Functions IF(condition, value if true, value if false) A logical function that returns one value if the condition is true and another if the condition is false. AND(condition 1, condition 2, . . . ) A logical function that returns TRUE if all conditions are true and FALSE if not. OR(condition 1, condition 2, . . . ) A logical function that returns TRUE if any condition is true and FALSE if not. VLOOKUP(lookup_value, table_array, col_index_num, [range lookup]) Looks up a value in the leftmost column of a table (specified by the table_array) and returns a value in the same row from a column you specify (col_ index_num). HLOOKUP(lookup_value, table_array, row_index_num, [range lookup]) Looks up a value in the top row of a table and returns a value in the same column from a row you specify. INDEX(array, row_num, col_num) Returns a value or reference of the cell at the intersection of a particular row and column in a given range. MATCH(lookup_value, lookup_array, match_type) Returns the relative position of an item in an array that matches a specified value in a specified order. CHOOSE(index_num, value1, value2, . . . ) Returns a value from a list based on the position in the list, specified by index_num. DSUM(database, field, criteria) Sums values in the field column from a specified database that includes the column labels, using a criteria range that specifies the
records you want to sum using the same format as the Advanced Filter. Other useful database functions include DAVERAGE and DCOUNT. Excel Techniques Create a range name (Example 2.2): There are three options. 1. Select a cell; in the Name box, type the name and press Enter. 2. Select a cell; click Define Name on the Formulas tab and complete the dialog. 3. Highlight a range with text labels immediately to the right or left of the cell or range, or in the row immediately above or below the cell or range. On the Formulas tab, choose Create from Selection and complete the dialog. Create an Excel table (Example 2.3): Select the range of the data, including headers. Click Table from the Tables group on the Insert tab and make sure that the box for My Table Has Headers is checked. Sort data (Example 2.5): Select a single cell in the column you want to sort on, and in the Data tab, click the “AZ down arrow” button to sort from smallest to largest or the “AZ up arrow” button to sort from largest to smallest. Click the Sort button to specify criteria for more advanced sorting capabilities. Filter data (Example 2.7): Select any cell within the database. Then, from the Excel Data tab, click on Filter. Choose the filtering criteria.
Chapter 2 Database Analytics
Using the Advanced Filter (Example 2.9): Copy the headers from the database to an open location in the worksheet. Under the headers, specify the criteria that you want to use to filter the data. Multiple criteria in the same row are logically joined by “and,” while criteria in rows are joined by “or.” In general, criteria for numerical values can include =, 7 , 6 , 6 7 (not equal), 7 =, or 6 =. Choose Advanced next to the Filter in the Data tab. In the dialog box, enter the list range of the database and the criteria range. Apply Data Validation (Example 2.17): Select the cell range for which data validation will be applied, and then choose Data Validation from the Data Tools group on the Data tab in Excel 2016, or select Validation from the Data menu on a Mac. Complete the dialog. Use Form Controls (Examples 2.19 and 2.20): Activate the Developer tab on the ribbon. To insert a form control, click the Insert button in the Controls tab under the Developer menu, click on the control you want to use, and then click within your
105
orksheet. Right click the form control and select w Format Control and specify the parameters. Create a PivotTable (Example 2.21): Select any cell in the database and choose PivotTable from the Tables group under the Insert tab and follow the steps of the wizard. Insert a Slicer in a PivotTable (Example 2.22): Click on the PivotTable and choose Insert Slicer from the Analyze tab in the PivotTable Tools ribbon in Windows, or from the PivotTable Analyze tab on a Mac. StatCrunch StatCrunch provides the ability to easily create surveys for collecting data. This is useful for creating your own data sets and databases for class assignments. You can find video tutorials and stepbystep procedures with examples at https://www.statcrunch.com/5.0/example.php. We suggest that you first view the tutorials Getting started with StatCrunch and Working with StatCrunch sessions in the Basics section. Scroll down to the Surveys section and click on Creating surveys and Administering surveys and analyzing the results.
PROBLEMS AND EXERCISES Data Sets and Databases
Data Queries: Tables, Sorting, and Filtering
1. The Excel file MBA Motivation and Salary Expecta
4. Convert the Hospital Claim database to an Excel
tions provides data on students’ preMBA salary and postMBA salary expectations. Define range names for each of these ranges and then use the range names in formulas to find the expected salary increase for each of the respondents.
table. Use the techniques described in Example 2.4 to find: a. the total number of claims. b. the total number of outpatient claims. c. the total number of claims in January.
2. The Excel file Airbnb Samples provides 15 sample
room rates for five randomly selected Airbnb accommodations. Define range names for each of the samples and then use these range names in formulas to find the average value for each sample. 3. The Financial Forecasts database shows estimated
variable and fixed expenses of a firm for the first nine months of the coming year. Define range names for each month and the type of expense. Then use the range names in formulas to find the total for each month and type of expense.
5. Convert the Order Specifications database to an
Excel table. Use the techniques described in Example 2.4 to find: a. the total cost of all orders. b. the total cost of all orders placed with Multipurpose Fasteners. c. the total quantity of airframe bolts purchased. 6. The Excel file Stationary provides the sales data for a
stationary company, and each row represents an order
106
Chapter 2 Database Analytics
for its products. Convert these data into an Excel table and filter the central region for those that have sold more than 50 units. What is the distribution of their order? 7. The Excel file Global Industry Default and Rating
summarizes financial information for 32 European firms, in the manufacturing and services industries, and their perceived risk of default. Convert these data into an Excel table. Use tablebased calculations to find the average credit score, average debt, and average equity for companies with a risk of default, and also for those without a risk of default. Does there appear to be a difference between firms with and without a risk of default? 8. Open the Excel file Store and Regional Sales Data
base. Sort the data by units sold, high to low. 9. In the Excel file Automobile Quality Ranking, search
for and add a new column for the country of origin for each brand. Then sort the data from the lowest to highest by number of problems per 100 vehicles in Excel. What conclusions can you reach? 10. In the Purchase Orders database, conduct a Pareto
analysis of the cost per order data. What conclusions can you reach? 11. Conduct a Pareto analysis for the number of minutes
that flights are late in the Global Airline Data Excel file. Interpret the results. 12. Use Excel’s filtering capability in the Car database to
extract a. all Japanese cars. b. all cars having more than 400 horsepower. c. all cars having less than 8 cylinders and weighing more than 3,500 kilograms. 13. In the Sales Transactions database, use Excel’s filter
ing capability to extract a. all orders that used PayPal. b. all orders under $100. c. all orders that were over $100 and used a credit card. 14. Filter the data in the Cereal database to obtain the
cereals that are of type C. What are the average calories and total sales for this type of cereal? 15. Use the Excel file Credit Risk Data that was introduced
in this chapter to perform the following activities:
a. Compute the combined checking and savings
account balance for each record in the Base Data worksheet. Then sort the records by the number of months as a customer of the bank. From examining the data, does it appear that customers with a longer association with the bank have more assets? b. Apply Pareto analysis to draw conclusions about the combined amount of money in checking and savings accounts. c. Use Excel’s filtering capability to extract all records for newcar loans. How many individuals with newcar loans are single, married, and divorced? d. Use Excel’s filtering capability to extract all records for individuals employed less than 12 months. Can you draw any conclusions about the credit risk associated with these individuals? 16. Apply the Advanced Filter to the Credit Risk Data
(Base Data worksheet) to find the following: a. All new car loans obtained for single females. b. All furniture loans obtained for single individuals who rent. c. All education loans obtained by unskilled workers. d. All used car loans obtained by individuals who are employed 12 months or less. e. All car loans (new or used) obtained by males who have been employed at least 36 months. 17. In the Credit Risk Data file, use database functions
to find the average amount of savings for each of the situations listed in Problem 16. 18. For the Bicycle Inventory database, use database
functions to find the following: a. The total number of leisure bicycles on hand. b. The average purchase cost and selling price of
road bicycles. Logical Functions 19. Modify the data in the Base Data worksheet in the
Excel file Credit Risk Data using IF functions to include new columns, classifying the checking and savings account balances as low if the balance is less than $250, medium if at least $250 but less than $2000, and high otherwise.
Chapter 2 Database Analytics
20. The Excel file President’s Inn Guest Database pro
vides a list of customers, rooms they occupied, arrival and departure dates, number of occupants, and daily rate for a small bedandbreakfast inn during one month.7 Room rates include breakfast and are the same for one or two guests; however, any additional guests must pay an extra $20 per person per day for breakfast. Parties staying for seven days or more receive a 10% discount on the room rate as well as any additional breakfast fees. Modify the spreadsheet to calculate the number of days that each party stayed at the inn and the total revenue for the length of stay.
107
to that movie by the database. If they are the same, the database assigns the number of the duplicate. An example is shown in the following table. The first four people watched the same movie, so the codes are assigned from 1 to 4, and so on. Explain how to assign the correct code using an IF function. A
B
1
Type of Movie Repeat Code
2
Action
1
3
Action
2
4
Action
3
5
Action
4
6
Comedy
1
7
Comedy
2
8
Comedy
3
9
Drama
1
10
Drama
2
22. For the Cereal database, write a logical IF function
11
Drama
3
that enters “Review” in column M if the rating is more than 45 and the sales is less than 150; if not, then enter a blank.
12
Drama
4
13
Drama
5
14
Drama
6
23. A survey was conducted to identify the type of mov
15
Drama
7
16
Drama
8
21. Figure 2.41 shows the 2016 U.S. federal income
tax rates for the four different types of filing status. Suppose an individual is single. Write one logical IF function that will determine the correct tax rate for any income that is entered in cell B3 of a spreadsheet. Implement this on a spreadsheet and verify its correctness.
ies watched by 15 people in the past two weeks. If the type of movie watched by one person is different from the others, then the code 1 will be assigned
Figure 2.41 2016 Federal Income Tax Rates for Problem 21
1. Single
2. Married filing jointly or qualifying widow/widower
3. Married filing separately
4. Head of household
Up to $9,275
Up to $18,550
Up to $9,275
Up to $13,250
10%
$9,276 to $37,650
$18,551 to $75,300
$9,276 to $37,650
$13,251 to $50,400
15%
$37,651 to $91,150
$75,301 to $151,900
$37,651 to $75,950
$50,401 to $130,150
25%
$91,151 to $190,150
$151,901 to $231,450
$75,951 to $115,725
$130,151 to $210,800
28%
$190,151 to $413,350
$231,451 to $413,350
$115,726 to $206,675
$210,801 to $413,350
33%
$413,351 to $415,050
$413,351 to $466,950
$206,676 to $233,475
$413,351 to $441,000
35%
$415,051 or more
$466,951 or more
$233,476 or more
$441,001 or more
7
Tax rate
39.60%
Based on Kenneth C. Laudon and Jane P. Laudon, Essentials of Management Information Systems, 9th ed. (Upper Saddle River, NJ: Prentice Hall, 2011).
108
Chapter 2 Database Analytics
Lookup Functions for Database Queries 24. A shirt manufacturer needs to supply invoices to cus
tomers based on data from the Shirt Production Excel file, which indicates the costs (in £) and sizes. Modify the file to use the VLOOKUP function to compute the customer invoice when product specifications, such as type, size, or quantity ordered, are entered. 25. The Excel file S&P 500 provides open, high, low,
and close values for the S&P 500 index over a period of time. a. Enter any date (using the format month/day/year) within the range of the data in cell G2. Use a MATCH function in cell G3 to find the row in the database that corresponds to this date. b. Write an INDEX function in cell G4 that finds the closing value on this date using your answer to part a. c. Combine the MATCH and INDEX functions in parts a and b into one function that displays the closing value in cell G5. 26. Enhance the Sales Transactions database file to per
form a database query that finds the transaction code and amount associated with any customer ID input. Apply your results to customer ID 10029. 27. Enhance the Yield Rates database file to perform a data
base query that finds the rate associated with any date and term. (This is a challenging question because you will not be able to use the term headers in a MATCH function. Hint: Add a row that assigns a code from 1 to 11 for each term. Then use the code for your input value. You might also use a lookup function to convert the code back to the actual term as part of the output.) 28. For this exercise, use the Wikipedia database.
Use MATCH and/or INDEX functions to find the following: a. The row numbers corresponding to the first and last instance of name Affaire_RÃ©veillon in column B (be sure column B is sorted by order number). b. The date associated with the first instance of name Affaire_RÃ©veillon that you identified in part a. c. The total changes for Affaire_RÃ©veillon. Use the answers of parts a. and b. along with the SUM function to do this. In another words, you should use appropriate INDEX and MATCH functions within the SUM function to find the answer. Validate your result by applying the SUM function directly to the data in column E.
29. Use INDEX and MATCH functions to fill in a table
that extracts the amounts shipped between each pair of cities in the Excel file General Appliance Corporation, which shows the solution to an optimization model that finds the minimum cost and amounts shipped from two plants to four distribution centers. Your table should be set up as follows, and the formulas for the Amount Shipped column should reference the names in the From and To columns: From
To
Marietta
Cleveland
Marietta
Baltimore
Marietta
Chicago
Marietta
Phoenix
Minneapolis
Cleveland
Minneapolis
Baltimore
Minneapolis
Chicago
Minneapolis
Phoenix
Amount Shipped
30. In the Factvalue database, you need to find the
country with a birth rate of 29.43 and identify the country’s death rate. Write functions for a left table lookup to answer these queries. 31. An auditor found an expense receipt for $179.24 with
out a name. Write a left table lookup function that would identify the sales rep in the Travel Expenses database. 32. Use the CHOOSE function to develop a database
query for the Time Spent data that will complete the following table for any individual specified in the top left corner:
Individual
Household Personal Free and Family Care Care Sleep Eating Time
Time Spent
Each empty cell should contain only one CHOOSE function. Hint: Use other appropriate function to determine the row associated with the specified individual. Excel Template Design 33. Suppose that an electricity supplier in your country
offers specific rates to domestic consumers. For the first 2,000 units of electricity consumed, it charges €0.14 per unit; and if more than 2,000 units are consumed, it charges €0.11 per unit consumed. Develop an Excel template using the VLOOKUP function to find the costs associated with any amount of consumption and compute the total cost of the consumption.
Chapter 2 Database Analytics
34. Develop an Excel template for computing the IQ
scores obtained by males and females based on the brain size (in count obtained from MRI scans, count/10,000), height (in inches) and weight (in pounds). Use the following formulas. For a male, the IQ score is determined by 110.6 + 2.16 * brain size  3.73 * height + 0.051 * weight For a female, the IQ score is determined by 101.45 + 3.59 * brain size + 2.83 * height  0.006 * weight Use form controls to input each individual’s brain size, height, weight, and gender. 35. A college graduate who is living in Edinburgh, the
United Kingdom, expects to earn a monthly salary of £598. The Excel file UK Cost of Living shows the comparative salaries in other cities and the adjustments for living expenses. Develop an Excel template that allows the user to enter current annual expenses (in Edinburgh) for cost of living, rent, groceries, and dining and compute the corresponding expenses for any selected city and the monthly wages surplus for both Edinburgh and the selected city. The following table is an example of what the output should look like: Edinburgh
Newcastle
£598.00
£501.00
Cost of Living
£79.39
£90.50
Rent
£35.55
£37.33
Groceries
£63.89
£70.92
Dining
£86.39
£98.48
Total
£265.22
£297.23
Wage Surplus
£332.78
£203.77
Monthly Wages Monthly Expenses
109
Use a list box form control to select the city and sliders to input current monthly expenses. 36. The Excel file Payroll Data provides hourly salaries
for a group of employees. Create an Excel template that allows the user to select an employee by employee ID, enter the number of regular hours and overtime hours worked, and display a payroll summary with the employee name, gross pay, federal tax, state tax, Social Security, Medicare withholding deductions, and net pay. Assume that the federal tax rate is 11%, the state tax rate is 2.385%, Social Security withholding is 6.2%, and Medicare withholding is 1.45%. Use a form control to select the employee ID. (Hint: Use the CONCATENATE function to join the first and last name in your template. See Chapter 1 Appendix A1 for a discussion of this function.) PivotTables 37. Construct PivotTables showing the counts of region
versus social media platform marketshare, and tablet versus mobile device usage in Asia in the Excel file Social Media Marketshare. What might you conclude from this analysis? 38. Use PivotTables to find the number of loans by dif
ferent purposes, marital status, and credit risk in the Excel file Credit Risk Data (Base Data worksheet). 39. Use PivotTables to find the number of sales transac
tions by product and region, total amount of revenue by region, and total amount of revenue by region and product in the Sales Transactions database. 40. The Excel file Retail Survey provides data about cus
tomers’ preferences for denim jeans. Use PivotTables and slicers to draw conclusions about how preferences differ by gender and age group and summarize your results in a brief memo.
CASE: PEOPLE’S CHOICE BANK The People’s Choice Bank is a small community bank that has three local branches, in Blue Ash, Delhi, and Anderson Hills. The Excel file Peoples Choice Bank is a database of major account transactions for the month of August. 1. Note that it is difficult to determine whether each transaction in the Amount column represents a positive cash inflow or a negative cash outflow without information
in other columns. Modify the database as appropriate to more easily analyze the data. 2. Suppose that you have been asked to prepare a summary of these transactions that would be clear and meaningful to the bank’s president. Use PivotTables (that are welldesigned and properly formatted) to provide a summary of the key information that the president and his direct reports would want to know. Justify and explain your reasoning.
110
Chapter 2 Database Analytics
CASE: DROUT ADVERTISING RESEARCH PROJECT8 Jamie Drout is interested in perceptions of gender stereotypes within beauty product advertising, which includes soap, deodorant, shampoo, conditioner, lotion, perfume, cologne, makeup, chemical hair color, razors, skin care, feminine care, and salon services, as well as the perceived benefits of empowerment advertising. Gender stereotypes specifically use cultural perceptions of what constitutes an attractive, acceptable, and desirable man or woman, frequently exploiting specific gender roles, and are commonly employed in advertisements for beauty products. Women are represented as delicately feminine, strikingly beautiful, and physically flawless, occupying small amounts of physical space that generally exploit their sexuality; men are represented as strong and masculine, with chiseled physical bodies, occupying large amounts of physical space to maintain their masculinity and power. In contrast, empowerment advertising strategies negate gender stereotypes and visually communicate the unique differences in each individual. In empowerment advertising, men and women are to represent the diversity in beauty, body type, and levels of perceived femininity and masculinity. Jamie’s project is focused on understanding consumer perceptions of these advertising strategies. Jamie conducted a survey using the following questionnaire: 1. What is your gender? Male Female
$10,000 to 6 $20,000 $20,000 to 6 $30,000 $30,000 to 6 $40,000 $40,000 to 6 $50,000 $50,000 to 6 $60,000 $60,000 to 6 $70,000 $70,000 to 6 $80,000 $80,000 to 6 $90,000 $90,000 to 6 $110,000 $110,000 to 6 $130,000 $130,000 to 6 $150,000 $150,000 or More 5. On average, how much do you pay for beauty and hygiene products or services per year? Include references to the following products: soap, deodorant, shampoo, conditioner, lotion, perfume, cologne, makeup, chemical hair color, razors, skin care, feminine care, and salon services. 6. On average, how many beauty and hygiene advertisements, if at all, do you think you view or hear per day? Include references to the following advertisements: television, billboard, Internet, radio, newspaper, magazine, and direct mail. 7. On average, how many of those advertisements, if any, specifically subscribe to gender roles and stereotypes?
2. What is your age? 3. What is the highest level of education you have completed? Some High School Classes High School Diploma Some Undergraduate Courses Associate Degree Bachelor Degree Master Degree J.D. M.D. Doctorate Degree 4. What is your annual income? $0 to 6 $10,000 8
8. On the following scale, what role, if any, do these advertisements have in reinforcing specific gender stereotypes? Drastic Influential Limited Trivial None 9. To what extent do you agree that empowerment advertising, which explicitly communicates the unique differences in each individual, would help transform cultural gender stereotypes? Strongly agree Agree Somewhat agree Neutral
The author expresses appreciation to Jamie Drout for providing this original material from her class project as the basis for this case.
Chapter 2 Database Analytics
Somewhat disagree Disagree Strongly disagree 10. On average, what percentage of advertisements that you view or hear per day currently utilize empowerment advertising? Assignment: Jamie received 105 responses, which are organized as a database in the Excel file Drout Advertising Survey. 1. Explain how the data and subsequent analysis using business analytics might lead to a better understanding of stereotype versus empowerment advertising.
111
Specifically, state some of the key insights that you would hope to answer by analyzing the data. 2. Create some PivotTables to draw some initial insights from the data. An important aspect of business analytics is good communication. Write up your answers to this case formally in a wellwritten report as if you were a consultant to Ms. Drout. This case will continue in Chapters 3, 4, 6, and 7, and you will be asked to use a variety of descriptive analytics tools to analyze the data and interpret the results. Your instructor may ask you to add your insights to the report, culminating in a complete project report that fully analyzes the data and draws appropriate conclusions.
This page intentionally left blank
CHAPTER
3
Data Visualization
ESB Professional/Shutterstock
LEARNING OBJECTIVES ■■ ■■
After studying this chapter, you will be able to:
Create Microsoft Excel charts. Determine the appropriate chart to visualize different types of data.
■■
■■
Apply data bars, color scales, icon sets, and sparklines to create other types of visualizations. Develop useful dashboards for communicating data and information.
114
Chapter 3 Data Visualization
Making sense of large quantities of disparate data is necessary not only for gaining competitive advantage in today’s business environment, but also for surviving in it. Converting data into information to understand past and current performance is the core of descriptive analytics and is vital to making good business decisions. The old adage “A picture is worth 1,000 words” is probably truer in today’s information–rich environment than ever before. Data visualization is the process of displaying data (often in large quantities) in a meaningful fashion to provide insights that will support better decisions. Researchers have observed that data visualization improves decision making, provides managers with better analysis capabilities that reduce reliance on IT professionals, and improves collaboration and information sharing. In your career, you will most likely use data visualization extensively as an analysis tool and to communicate data and information to others. For example, if you work in finance, you can use data visualization to track revenues, costs, and profits over time; to compare performance between years or among different departments; and to track budget performance. In marketing, you can use data visualization to show trends in customer satisfaction, compare sales among different regions, and show the impact of advertising strategies. In operations, you might illustrate the performance of different facilities, product quality, call volumes in a technical support department, or supply chain metrics such as late deliveries. In this chapter, we will illustrate how to construct and use Excel charts and other Excel visualization tools and how to build dashboards that summarize and communicate key information visually.
The Value of Data Visualization Raw data are important, particularly when one needs to identify accurate values or compare individual numbers. However, it is quite difficult to identify trends and patterns, find exceptions, or compare groups of data in tabular form. The human brain does a surprisingly good job of processing visual information—if presented in an effective way. Visualizing data provides a way of communicating data at all levels of a business and can reveal surprising patterns and relationships, thereby providing important insights for making decisions. It also helps users to more quickly understand and interpret data, and helps analysts choose the most appropriate data analysis tool. For many unique and intriguing examples of data visualization, visit the Data Visualization Gallery at the U.S. Census Bureau Web site, www.census.gov/dataviz/. In addition to descriptive analytics, data visualization is important for predictive and prescriptive analytics as well. For example, recall the chart predicting new product sales over time, shown in Figure 1.2 in Chapter 1. This graph conveys the concept much more
Chapter 3 Data Visualization
115
easily than would either a verbal description or mathematical model. Visualizing a pattern also helps analysts select the most appropriate mathematical function to model the phenomenon. Complex prescriptive models often yield complex results. Visualizing the results often helps in understanding and gaining insight about model output and solutions.
EXAMPL E 3.1
Tabular Versus Visual Data Analysis
Figure 3.1 shows the data in the Excel file Monthly Product Sales. We can use the data to determine exactly how many units of a certain product were sold in a particular month, or to compare one month to another. For example, we see that sales of product A dropped in February, specifically by 6.7% (computed by the Excel formula = 1  B3/B2). Beyond such calculations, however, it is difficult to draw big–picture
▶▶ Figure
3.1
Monthly Product Sales Data
▶▶ Figure
3.2
Visualization of Monthly Product Sales Data
conclusions. Figure 3.2 displays a chart of monthly sales for each product. We can easily compare overall sales of different products (product C sells the least, for example) and identify trends (sales of product D are increasing), other patterns (sales of product C are relatively stable, while sales of product B fluctuate more over time), and exceptions (product E’s sales fell considerably are September).
116
Chapter 3 Data Visualization
ANALYTICS IN PRACTICE: Data Visualization for the New York City
Police Department’s Domain Awareness System1
The New York City Police Department (NYPD), the largest state or local police force in the United States, is charged with securing New York City from crime and terrorism. The department had accumulated a tremendous amount of information but had limited means of sharing it among its officers. Much of the information was available only to officers in the precinct house with permission to access standalone software applications, with little analytics or data visualization techniques to give the officers any insight. In 2008, the NYPD began developing a new system, the Domain Awareness System (DAS), which is a citywide network of sensors, databases, devices, software, and infrastructure that informs decision making by delivering analytics and tailored information to officers’ smartphones and precinct desktops. The NYPD has used the system to employ a unique combination of analytics and information technology.
A key feature of the system is data visualization. Prior to the DAS, reports simply listed data in numerical tables. The only analysis presented might be a yeartoyear percentage change. Using it to pick out geographic clusters or potential patterns was virtually impossible. Today, information is presented in an interactive form. If a user clicks on a number, DAS brings up all the records included in that number and marks them on the map. The NYPD also constructed a data visualization engine to enable the user to explore trends and patterns in the statistics. Bar and pie charts of categorical data and line charts of temporal data are available with the press of a button. Through improving the efficiency of the NYPD’s staff, DAS has generated estimated savings of $50 million per year. Most importantly, the NYPD has used it to combat terrorism and improve its crimefighting effectiveness. Since DAS was deployed departmentwide in 2013, the overall crime index in the city has fallen by six percent.
Tools and Software for Data Visualization Data visualization ranges from simple Excel charts to more advanced interactive tools and software that allow users to easily view and manipulate data with a few clicks, not only on computers, but also on iPads and other devices. In this chapter, we discuss basic tools available in Excel. Commercial software packages such as Tableau, QlikView, and SAS Visual Analytics offer more powerful tools, especially for applications involving big data. In particular, we suggest that you look at the capabilities of Tableau (www.tableau.com), which we describe in the Appendix to this chapter. Tableau is easy to use and offers a free trial. CHECK YOUR UNDERSTANDING 1. Explain the pros and cons of tabular versus visual data analysis. 2. How is data visualization used in descriptive, predictive, and prescriptive analytics?
Creating Charts in Microsoft Excel Microsoft Excel provides a comprehensive charting capability with many features. With a little experimentation, you can create professional charts for business analyses and presentations. These include vertical and horizontal bar charts, line charts, pie charts, area charts, 1
E. S. Levine, Jessica Tisch, Anthony Tasso, and Michael Joya, “The New York City Police Department’s Domain Awareness System,” Interfaces, Vol. 47, No. 1, January–February 2017, pp. 70–84, © 2017 INFORMS.
Chapter 3 Data Visualization
117
scatter plots, and many other special types of charts. We will not guide you through every application but do provide some guidance for new procedures as appropriate. Certain charts work better for certain types of data, and using the wrong chart can make the data difficult for the user to interpret and understand. While Excel offers many ways to make charts unique and fancy, naive users often focus more on the attentiongrabbing aspects of charts rather than their effectiveness in displaying information. So we recommend that you keep charts simple, and avoid such bells and whistles as 3D bars, cylinders, cones, and so on. We highly recommend books written by Stephen Few, such as Show Me the Numbers (Oakland, CA: Analytics Press, 2004), for additional guidance in developing effective data visualizations. To create a chart in Excel, it is best to first highlight the range of the data you wish to chart. The Excel Help files provide guidance on formatting your data for a particular type of chart. Click the Insert tab in the Excel ribbon (Figure 3.3; the Mac ribbon is similar). Click the chart type, and then click a chart subtype that you want to use. Once a basic chart is created, you may use the options in the Design (Chart Design in Mac) and Format tabs to customize your chart (Figure 3.4). In the Design tab, you can change the type of chart, data included in the chart, chart layout, and styles. The Format tab provides various formatting options. You may also customize charts easily by right clicking on elements of the chart or by using the Quick Layout options within the Design tab. You should realize that up to 10% of the male population are affected by color blindness, making it difficult to distinguish between different color variations. Although we generally display charts using Excel’s default colors, which often, unfortunately, use red, experts suggest using blueorange palettes. We suggest that you be aware of this for professional and commercial applications.
Column and Bar Charts Excel distinguishes between vertical and horizontal bar charts, calling the former column charts and the latter bar charts. A clustered column chart compares values across categories using vertical rectangles; a stacked column chart displays the contribution of each value to the total by stacking the rectangles; and a 100% stacked column chart compares the percentage that each value contributes to a total. Column and bar charts are useful for comparing categorical or ordinal data, for illustrating differences between sets of values, and for showing proportions or percentages of a whole.
▲▲ Figure
3.3
Excel Insert Ribbon for Windows
▲▲ Figure
3.4
Excel Chart Design Ribbon for Windows
118
Chapter 3 Data Visualization
EXAMPLE 3.2
Creating Column Charts
The Excel file EEO Employment Report provides data on the number of employees in different categories broken down by racial/ethnic group and gender (Figure 3.5). We will construct a simple column chart for the various employment categories for all employees. First, highlight the range C3:K6, which includes the headings and data for each category. Click on the Column Chart button found in the Insert tab, and then on the first chart type in the list (a clustered column chart). To add a title, click on the Add Chart Elements button in the Design tab ribbon. Click on “Chart Title” in the chart
and change it to “Alabama Employment.” The names of the data series can be changed by clicking on the Select Data button in the Data group of the Design tab. In the Select Data Source dialog (see Figure 3.6), click on “Series1” and then the Edit button. Enter the name of the data series, in this case “All Employees.” Change the names of the other data series to “Men” and “Women” in a similar fashion. You can also change the order in which the data series are displayed on the chart using the up and down buttons. The final chart is shown in Figure 3.7.
Be cautious when changing the scale of the numerical axis. The heights or lengths of the bars only accurately reflect the data values if the axis starts at zero. If not, the relative sizes can paint a misleading picture of the true relative values of the data.
Data Label and Data Table Chart Options Excel provides options for including the numerical data on which charts are based within the charts. Data labels can be added to chart elements to show the actual value of bars,
▲▲ Figure
3.5
Portion of EEO Employment Report Data ▶▶ Figure
3.6
Select Data Source Dialog
▶▶ Figure
Chapter 3 Data Visualization
119
3.7
Column Chart for Alabama Employment Data
for example. Data tables can also be added; these are usually better than data labels, which can get quite messy. Both can be added from the Add Chart Element button in the Chart Tools Design tab, or from the Quick Layout button, which provides standard design options. Figure 3.8 shows a data table added to the Alabama Employment chart. You can see that the data table provides useful additional information to improve the visualization.
Line Charts Line charts provide a useful means for displaying data over time, as Example 3.3 illustrates. You may plot multiple data series in line charts; however, they can be difficult to interpret if the magnitude of the data values differs greatly. In that case, it would be advisable to create separate charts for each data series.
▲▲ Figure
3.8
Alternate Column Chart Format with a Data Table
120
Chapter 3 Data Visualization
EXAMPLE 3.3
A Line Chart for China Export Data
Figure 3.9 shows a line chart giving the amount of U.S. exports to China in billions of dollars from the Excel file China Trade Data. The chart clearly shows a significant rise
in exports starting in the year 2000, which began to level off around 2008 and then show a sharp increase in subsequent years.
Pie Charts For many types of data, we are interested in understanding the relative proportion of each data source to the total. A pie chart displays this by partitioning a circle into pieshaped areas showing the relative proportions. Example 3.4 provides one application. EXAMPLE 3.4
A Pie Chart for Census Data
Consider the marital status of individuals in the U.S. population in the Excel file Census Education Data, a portion of which is shown in Figure 3.10. To show the relative proportion in each category, we can use a pie chart, as shown in
▶▶ Figure
3.9
Line Chart for China Trade Data
▶▶ Figure
3.10
Portion of Census Education Data
Figure 3.11. This chart uses a layout option that shows the labels associated with the data as well as the actual proportions as percentages. A different layout that shows the values and/or proportions can also be chosen.
Chapter 3 Data Visualization
▶▶ Figure
121
3.11
Pie Chart for Marital Status: Not a High School Grad
Data visualization professionals don’t recommend using pie charts. For example, contrast the pie chart in Figure 3.11 with the column chart in Figure 3.12 for the same data. In the pie chart, it is difficult to compare the relative sizes of areas; however, the bars in the column chart can easily be compared to determine relative ratios of the data. If you do use pie charts, restrict them to small numbers of categories, always ensure that the numbers add to 100%, and use labels to display the group names and actual percentages. Avoid threedimensional (3D) pie charts—especially those that are rotated—and keep them simple.
Area Charts An area chart combines the features of a pie chart with those of line charts. Area charts present more information than pie or line charts alone but may clutter the observer’s mind with too many details if too many data series are used; thus, they should be used with care. EXAMPL E 3.5
An Area Chart for Energy Consumption
Figure 3.13 displays total energy consumption (billions Btu) and consumption of fossil fuels from the Excel file Energy Production & Consumption. This chart shows that although total energy consumption has grown since 1949, the relative
▶▶ Figure
3.12
Alternative Column Chart for Marital Status: Not a High School Grad
proportion of fossil fuel consumption has remained generally consistent at about half of the total, indicating that alternative energy sources have not replaced a significant portion of fossil fuel consumption.
122
Chapter 3 Data Visualization
▶▶ Figure
3.13
Area Chart for Energy Consumption
Scatter Charts and Orbit Charts Scatter charts show the relationship between two variables. To construct a scatter chart, we need observations that consist of pairs of variables. For example, students in a class might have grades for both a midterm and a final exam. A scatter chart would show whether high or low grades on the midterm correspond strongly to high or low grades on the final exam, or whether the relationship is weak or nonexistent. EXAMPLE 3.6
A Scatter Chart for Real Estate Data
Figure 3.14 shows a scatter chart of house size (in square feet) versus the home market value from the Excel file Home
Market Value. The data clearly suggest that higher market values are associated with larger homes.
An orbit chart is a scatter chart in which the points are connected in sequence, such as over time. Orbit charts show the “path” that the data take over time, often showing some unusual patterns that can provide unique insights. You can construct an orbit chart by creating a scatter chart with smooth lines and markers from the scatter chart options. Figure 3.15 shows an example from the Gasoline Sales Excel file using the first 10 weeks of data. ▶▶ Figure
3.14
Scatter Chart of House Size Versus Market Value
Chapter 3 Data Visualization
▶▶ Figure
123
3.15
Orbit Chart
Bubble Charts A bubble chart is a type of scatter chart in which the size of the data marker corresponds to the value of a third variable; consequently, it is a way to plot three variables in two dimensions.
EXAMPL E 3.7
A Bubble Chart for Comparing Stock Characteristics
Figure 3.16 shows a bubble chart for displaying price, P/E (price/earnings) ratio, and market capitalization for five different stocks on one particular day from the Excel file Stock
▶▶ Figure
3.16
Bubble Chart for Stock Comparisons
Comparisons. The position on the chart shows the price and P/E; the size of the bubble represents the market cap in billions of dollars.
124
Chapter 3 Data Visualization
Combination Charts Often, we wish to display multiple data series on the same chart using different chart types. Excel 2016 for Windows provides a Combo Chart option for constructing such a combination chart; in Excel 2016 for Mac, it must be done manually. We can also plot a second data series on a secondary axis; this is particularly useful when the scales differ greatly.
EXAMPLE 3.8
Creating a Combination Chart and Secondary Axis
Figure 3.17 shows data that have been added to the Monthly Product Sales data–a sales goal for product E and the percent of the goal that was actually achieved. We will first construct a chart that shows the sales of product E as compared to the monthly goals. In Excel 2016 for Windows, select the data in columns G and H, and from the Charts options in the Insert ribbon, select Insert Combo Chart. Figure 3.18 shows the result. On a Mac, create a standard column chart using both data series. Then right click the product E goal data series, select Change Chart Type, and select Line.
▶▶ Figure
3.17
Monthly Product Sales with Additional Data
▶▶ Figure
3.18
Combination Chart
Next, to plot product E sales and the percent of the goal achieved, select the data in columns F and H (first select the data in column F, then hold the Ctrl key and select the data in column H). Choose a Combo Chart again. Because the percent of goal data series is very small in comparison to product sales, the line chart will look like it is on the xaxis. Carefully right click on this data series and select Format Data Series > Axis > Plot series on secondary axis. The scale for % Goal will be added on the secondary axis on the right of the chart, as in Figure 3.19.
Chapter 3 Data Visualization
▶▶ Figure
125
3.19
Combination Chart with Secondary Axis
Radar Charts Radar charts show multiple metrics on a spider web. This is a useful chart to compare survey data from one time period to another or to compare performance of different entities such as factories, companies, and so on using the same criteria. EXAMPL E 3.9
A Radar Chart for Survey Responses
Figure 3.20 shows average survey responses for six questions on a customer satisfaction survey. The radar chart compares
the average responses between the first and second quarters. You can easily see which responses increased or decreased.
Stock Charts A stock chart allows you to plot stock prices, such as daily high, low, and close values. It may also be used for scientific data such as temperature changes. We will explain how to create stock charts in Chapter 6 to visualize some statistical results, and again in C hapter 15 to visualize optimization results.
Charts from PivotTables If you click inside a PivotTable, you can easily insert a chart that visualizes the data in the PivotTable from the Insert tab using any of the recommended charts, or by choosing your ▶▶ Figure
3.20
Radar Chart
126
Chapter 3 Data Visualization
own. In Excel for Windows, this is called a PivotChart (see the Appendix to this chapter for further information about PivotCharts). You can do this on a Mac also; however, charts created from PivotTables will not have the Filter buttons as described in the Appendix.
Geographic Data Many applications of business analytics involve geographic data. For example, finding the best location for production and distribution facilities, analyzing regional sales performance, transporting raw materials and finished goods, and routing vehicles such as delivery trucks involve geographic data. In such problems, data mapping can help in a variety of ways. Visualizing geographic data can highlight key data relationships, identify trends, and uncover business opportunities. In addition, it can often help to spot data errors and help end users understand solutions, thus increasing the likelihood of acceptance of decision models. Companies like Nike use geographic data and information systems for visualizing where products are being distributed and how that relates to demographic and sales information. This information is vital to marketing strategies. The use of prescriptive analytic models in combination with data mapping was instrumental in the success of Procter & Gamble’s North American supply chain study, which saved the company in excess of $200 million dollars per year.2 We discuss this application in Chapter 14. Excel 2016 for Windows includes a geographic visualization tool called 3D Maps. We encourage you to explore this. Another excellent option is Tableau, which is described in the Appendix to this chapter, in which we will illustrate its geographic visualization capabilities. CHECK YOUR UNDERSTANDING 1. Summarize the most useful charts available in Excel and the types of applications for which they should be used. 2. What is the difference between a bar and a column chart? 3. Why don’t data visualization professionals recommend the use of pie charts? 4. How do you create a combination chart in Excel?
Other Excel Data Visualization Tools Microsoft Excel offers numerous other tools to help visualize data. These include data bars, color scales, icon sets, and sparklines. These options are part of Excel’s Conditional Formatting rules, which allow you to visualize different numerical values through the use of colors and symbols. Excel has a variety of standard templates to use, but you may also customize the rules to meet your own conditions and styles using the New Formatting Rule option. This allows you to format cells that only contain certain values, those that are above or below average, as well as other rules. We encourage you to experiment with these tools.
Data Bars Data bars display colored bars that are scaled to the magnitude of the data values (similar to a bar chart) but placed directly within the cells of a range. 2
J. Camm et al., “Blending OR/MS, Judgment and GIS: Restructuring P&G’s Supply Chain,” Interfaces, 27, 1 (1997): 128–142.
Chapter 3 Data Visualization
▶▶ Figure
127
3.21
Example of Data Bars
EXAMPL E 3.10
Data Visualization with Data Bars
Figure 3.21 shows data bars applied to the data in the Monthly Product Sales worksheet. Highlight the data in each column, click the Conditional Formatting button in the Styles group within the Home tab, select Data Bars, and choose the fill option and color. You may also display data bars without the data in the cells. A useful tip is to copy the data next to the original data to display the data bars along with the original data.
Figure 3.22 illustrates this for the total monthly sales. We first summed the monthly sales in column G, then copied these to column H. Then highlight the range of data in column H, click Conditional Formatting, choose Data Bars, and select More Rules. In the Edit Formatting Rule dialog, check the box Show Bar Only. If some data are negative, data bars will display them on the left side of a vertical axis, allowing you to clearly visualize both positive and negative values.
Color Scales Color scales shade cells based on their numerical value using a color palette. This is another option in the Conditional Formatting menu. EXAMPL E 3.11
Data Visualization with Color Scales
Figure 3.23 shows the use of a greenyellowred color scale for the monthly product sales, which highlights cells containing large values in green, small values in red, and middle values in yellow. The darker the green, the larger the value;
▶▶ Figure
3.22
Displaying Data Bars Outside of Data Cells
the darker the red, the smaller the value. For intermediate values, you can see that the colors blend together. This provides a quick way of identifying the largest and smallest productmonth sales values.
128
Chapter 3 Data Visualization
▶▶ Figure
3.23
Example of Color Scales
Colorcoding of quantitative data is commonly called a heatmap. Heatmaps are often used to display geographic data, such as population densities, or other socioeconomic metrics by nation, state, county, and so on.
Icon Sets Icon sets provide similar information as color scales using various symbols such as arrows (see Figure 3.24) or red/yellow/green stoplights. Many companies use green, yellow, and red stoplight symbols to show good, marginal, and poor performance, respectively, in business dashboards. The next example illustrates this, and also shows how to customize a conditional formatting rule. EXAMPLE 3.12
Data Visualization with Customized Icon Sets
Figure 3.25 shows a set of stoplight icons that code the monthly product sales for each product as green if they are in the top 20% of the data range, red if in the bottom 20%, and yellow if in between. Note that because of the relative differences in the magnitude of sales among products, we created a new rule for each column of data. Highlight the data in a column, click Conditional Formatting, select Icon Sets, and then select More Rules. In the Edit Formatting Rule dialog, change the default values of 67 and 33% to 80 and 20%, as shown in the figure. To understand how this ▶▶ Figure
3.24
Example of Icon Sets
works, look at product B. The minimum value is 3,024 and the maximum is 5,991. So 80% is 3024 + 0.8 × (5991 − 3024) = 5397.6. Thus, any cell value above this is coded as green. Similarly, 20% is 3024 + 0.2 × (5991 − 3024) = 3617.4; any cell value below this is coded as red. In the Edit Formatting Rule dialog, you may also change the rule to code the cells based on their actual values by changing Percent to Number from the dropdown box. For example, you could code all values for product B greater than or equal to 5,000 as green and all values below 4,000 as red.
Chapter 3 Data Visualization
▲▲ Figure
129
3.25
Creating Customized Icon Sets
Sparklines Sparklines are graphics that summarize a row or column of data in a single cell. Sparklines were introduced by Edward Tufte, a famous expert on visual presentation of data. He described sparklines as “dataintense, designsimple, wordsized graphics.” Excel has three types of sparklines: line, column, and win/loss. Line sparklines are useful for timeseries data, while column sparklines are more appropriate for categorical data. Win/loss sparklines are useful for data that move up or down over time. They are found in the Sparklines group within the Insert tab on the ribbon.
EXAMPL E 3.13
Examples of Sparklines
We will again use the Monthly Product Sales data. Figure 3.26 shows line sparklines in row 16 for each product. In column G, we display column sparklines, which are essentially small column charts. Generally, you need to expand the row or column widths to display them effectively. Notice, however, that the lengths of the bars are not scaled properly to the data; for example, in the first one, products D and E are roughly onethird the value of product E, yet the bars are not scaled correctly. Excel defaults to automatic scaling of the minimum and maximum values of the vertical axis. You can change this
and scale the bars properly by choosing Axis from Group in the Sparkline Tools Design tab. Figure 3.27 shows a modified worksheet in which we computed the percentage change from one month to the next for products A and B. The win loss sparklines in row 16 show the patterns of sales increases and decreases, suggesting that product A has a cyclical pattern while product B changed in a more random fashion. If you click on any cell containing a sparkline, the Sparkline Tools Design tab appears, allowing you to customize colors and other options.
130
Chapter 3 Data Visualization
▶▶ Figure
3.26
Line and Column Sparklines
▶▶ Figure
3.27
Win/Loss Sparklines
CHECK YOUR UNDERSTANDING 1. Explain the purpose of data bars, color scales, and icon sets. 2. What is a heatmap, and what are some typical applications? 3. How are sparklines different from standard Excel charts?
Chapter 3 Data Visualization
131
Dashboards Making data visible and accessible to employees at all levels is a hallmark of effective modern organizations. A dashboard is a visual representation of a set of key business measures. It is derived from the analogy of an automobile’s control panel, which displays speed, gasoline level, temperature, and so on. Dashboards provide important summaries of key business information to help manage a business process or function. For example, the Cincinnati Zoo (see the introduction in Chapter 1) uses hourly, daily, and yearly dashboards that show such metrics as attendance and types of admission, cities where visitors come from, and revenue at different food and retail locations. Dashboards are particularly useful for senior managers who don’t have the time to sift through large amounts of data and need a summary of the state of business performance during monthly or quarterly reviews. Dashboards might include tabular as well as visual data to allow managers to quickly locate key data. Figure 3.28 shows a simple dashboard for the product sales data in Figure 3.1. Dashboards often incorporate color scales or icon sets to quickly pinpoint areas of concern. This one displays the monthly sales for each product individually, sales of all products combined, total annual sales by product, a comparison of the last two months, and monthly percent changes by product. An effective dashboard should capture all the key information that the users need for making good decisions. Important business metrics are often called key performance indicators (KPIs). People tend to look at data at the top left first, so the most important charts should be positioned there. An important principle in dashboard design is to keep it simple—don’t clutter the dashboard with too much information or use formats (such as 3D charts) that don’t clearly convey information.
▲▲ Figure
3.28
Dashboard for Product Sales
CHECK YOUR UNDERSTANDING 1. What is a dashboard? 2. What are the key design principles for dashboards? 3. What is a key performance indicator (KPI)?
132
Chapter 3 Data Visualization
ANALYTICS IN PRACTICE: Driving Business Transformation with
IBM Business Analytics3
Founded in the 1930s and headquartered in Ballinger, Texas, Mueller is a leading retailer and manufacturer of preengineered metal buildings and metal roofing products. Today, the company sells its products directly to consumers all over the southwestern United States from 35 locations across Texas, New Mexico, Louisiana, and Oklahoma. Historically, Mueller saw itself first and foremost as a manufacturer; the retail aspects of the business were a secondary focus. However, in the early 2000s, the company decided to shift the focus of its strategy and become much more retailcentric—getting closer to its enduse customers and driving new business through a better understanding of their needs. To achieve its transformation objective, the company needed to communicate its retail strategy to employees across the organization. As Mark Lack, Manager of Strategy Analytics and Business Intelligence at Mueller, explains: “The transformation from pure manufacturing to retailled manufacturing required a more endcustomerfocused approach to sales. We wanted a way to track how successfully our sales teams across the country were adapting to this new strategy, and identify where improvements could be made.” To keep track of sales performance, Mueller worked with IBM to deploy IBM Cognos Business Intelligence. The IBM team helped Mueller apply technology to its balanced scorecard process for strategy management in Cognos Metric Studio. By using a common set of KPIs, Mueller can easily identify the strengths and weaknesses of all of its sales teams through sales performance analytics. “Using Metric Studio in Cognos Business Intelligence, we get a clear picture of each team’s strategy performance,” says Mark Lack. “Using sales performance insights from Cognos scorecards, we can identify teams that are hitting their targets, and determine the reasons for their success. We can then share this knowledge with underperforming teams, and demonstrate how they can change their way of working to meet their targets.” Instead of just trying to impose or enforce new ways of working, we are able to show sales teams exactly how they are contributing to the business, and explain what they
3
need to do to improve their metrics. It’s a much more effective way of driving the changes in behavior that are vital for business transformation.” Recently, IBM Business Analytics Software Services helped Mueller upgrade to IBM Cognos 10. With the new version in place, Mueller has started using a new feature called Business Insight to empower regional sales managers to track and improve the performance of their sales teams by creating their own personalized dashboards. “Static reports are a good starting point, but people don’t enjoy reading through pages of data to find the information they need,” comments Mark Lack. “The new version of Cognos gives us the ability to create customized interactive dashboards that give each user immediate insight into their own specific area of the business, and enable them to drill down into the raw data if they need to. It’s a much more intuitive and compelling way of using information.” Mueller now uses Cognos to investigate the reasons why some products sell better in certain areas, which of its products have the highest adoption rates, and which have the biggest margins. Using these insights, the company can adapt its strategy to ensure that it markets the right products to the right customers—increasing sales. By using IBM SPSS Modeler to mine enormous volumes of transactional data, the company aims to reveal patterns and trends that will help to predict future risks and opportunities, as well as uncover unseen problems and anomalies in its current operations. One initial project with IBM SPSS Modeler aims to help Mueller find ways to reduce its fuel costs. Using SPSS Modeler, the company is building a sophisticated statistical model that will automate the process of analyzing fuel transactions for hundreds of vehicles, drivers, and routes. “With SPSS Modeler, we will be able to determine the average fuel consumption for each vehicle on each route over the course of a week,” says Mark Lack. “SPSS will automatically flag up any deviations from the average consumption, and we then drill down to find the root cause. The IBM solution helps us to determine if higherthanusual fuel transactions are legitimate—for example, a driver covering extra miles—or the result of some other factor, such as fraud.”
“Mueller builds a customerfocused business,” IBM Software, Business Analytics, © IBM Corporation, 2013.
Chapter 3 Data Visualization
133
KEY TERMS Icon sets Key performance indicator (KPI) Line chart Orbit chart Pie chart Radar chart Scatter chart Sparklines Stock chart
Area chart Bar chart Bubble chart Color scales Column chart Combination chart Dashboard Data bars Data visualization Heatmap CHAPTER 3 TECHNOLOGY HELP Excel Techniques Creating a chart (Example 3.2): Highlight the range of the data to chart. Click the Insert tab in the Excel ribbon. Click the chart type, and then click a chart subtype that you want to use. Use the options in the Design (Chart Design on a Mac) and Format tabs within the Chart Tools tabs to customize the chart, or use the Quick Layout options. To add a title, click on the Add Chart Elements button in the Design tab ribbon. The names of the data series can be changed by c licking on the Select Data button in the Data group of the Design tab.
StatCrunch StatCrunch provides various ways to chart and visualize data, including charts that are difficult to implement in Excel and are not discussed in this chapter. You can find video tutorials with stepbystep procedures and Study Card examples at https://www.statcrunch.com/5.0/ example.php. We suggest that you first view the tutorials Getting started with StatCrunch and Working with StatCrunch sessions. The following tutorials are listed in the Graphs section on this Web page and explain how to create basic charts:
Creating combination charts (Example 3.8): To display multiple data series on the same chart u sing different chart types in Excel 2016 for Windows, select the data series in two columns, and from the Charts options in the Insert ribbon, select Insert Combo Chart. On a Mac, create a standard column chart using both data series. Then right click the data series you wish to change and select Change Chart Type.
■■ ■■ ■■ ■■ ■■ ■■
Display data bars, color scales, and icon sets (Examples 3.10–3.12): Highlight the data, click the Conditional Formatting button in the Styles group within the Home tab, and select Data Bars, Color Scales, or Icon Sets. Choose the display you want or select More Rules to customize it. You may also select from other predetermined rules from the Conditional Formatting menu.
■■ ■■ ■■ ■■
Displaying Sparklines (Example 3.13): Select Sparklines from the Insert tab. Choose Line, Column, or Win/Loss, and complete the dialog. Expand the height or width of the cells to improve the visualization.
Pie charts from raw data Pie charts with summary data Split and stacked bar plots Charting values across multiple columns Scatter plots Bubble plots
You can also find tutorials for additional charts and methods for customizing colors and styles: Dotplots Stem and leaf plots Boxplots with a group by column Painting/annotating graphs
Example: Loading a file Click the Data menu, choose Load, and select the option. Example: Create a Scatter Plot 1. Select the X column and Y column for the plot. 2. Enter an optional Where statement to specify the data rows to be included. 3. Colorcode points with an optional Group by column. 4. Click Compute! to produce the plot.
134
Chapter 3 Data Visualization
Example: Create a Pie Chart with Summary Data 1. Select the variable that contains the categories and the variable that contains the counts for the plot.
2. Enter an optional Where statement to specify the data rows to be included. 3. Click Compute! to produce the plot.
PROBLEMS AND EXERCISES Creating Charts in Microsoft Excel 1. Create clustered column and stacked column charts
for the preMBA and postMBA salary data in the Excel file MBA Motivation and Salary Expectations. Discuss which type of chart you feel is better to use to explain the information. 2. Create a line chart illustrating the number of AM
and PM travelers for each month in the Excel file Traveler. 3. Create a pie chart for all the given dates and a stock
chart for the high/low/close prices for March 2018 in the Stock Excel file. 4. The Excel file Expenditure contains data on a coun
try’s expenditure on various sports during a particular year. Create a pie chart showing the percentage breakdown of total expenditure. 5. Create a stacked area chart for the monthly average
sales of the different types of fuel in the Excel file Average Fuel Sales. What conclusion can you reach? 6. The Excel file Mortgage Installment provides the
monthly installment of a mortgage loan by different fixed interest rate (%) provided by different banks. Construct a scatter diagram to show the relationship between the monthly installment and the fixed interest rate (%).What conclusion can you reach? 7. The Excel file Monthly Visits provides the data for
monthly visitors visited a playground and the number of shows prepared by the playground organizer during the specific month. Create a scatter diagram to show the relationship between the number of visitors and the number of shows. What conclusion can you reach? 8. Create a bubble chart for the first five colleges in the
Excel file Colleges and Universities for which the xaxis is the top 10% HS, yaxis is acceptance rate, and bubbles represent the expenditures per student. 9. Construct a column chart for the data in the Excel file
Purchasing Rate to allow comparison of the rate for
each product with the date. Would any other charts be better for visually conveying this information? Why? 10. The Excel file Transportation provides sales data
about 10 different industries. Construct two stacked bar charts that will allow you to compare any differences due to time or age groups and draw any conclusions that you can. Suggest another type of chart that may be more appropriate. 11. Construct an appropriate chart to show the sales per
unit for the primary industry in the Excel file Primary Industry. 12. Construct an appropriate chart or charts to visualize
the information in the Budget Forecasting Excel file. Explain why you chose the chart(s) you used. 13. A survey was conduct for determine the number of
vegetarians and nonvegetarians between two different group of age of people. The survey results the followings Age Group
Vegetarians
Nonvegetarians
30 to 6 40
186
432
Ú 40
239
293
Convert the data into percentages for each age group. Then construct a chart of count and a chart of percentages. Interpret both the charts and discuss whether there is a difference between the two charts. Other Excel Data Visualization Tools 14. In the Excel file Banking Data, apply the following
data visualization tools: a. Use data bars to visualize the relative values of median home value. b. Use color scales to visualize the relative values of median household wealth. c. Use an icon set to show high, medium, and low average bank balances, where high is above $30,000, low is below $10,000, and medium is anywhere in between.
Chapter 3 Data Visualization
15. Apply three different colors of data bars to lunch, din
ner, and delivery sales in the Excel file Restaurant Sales to visualize the relative amounts of sales. Then sort the data by the day of the week beginning on Sunday. Compare the nonsorted data with the sorted data and comment on the information content of the visualizations. 16. For the Store and Regional Sales database, apply a
fourtrafficlight icon set to visualize the distribution of the number of units sold for each store, where green corresponds to at least 30 units sold, yellow to at least 20 but less than 30, red to at least 10 but less than 20, and black to below 10. 17. For the Excel file Closing Stock Prices, a. apply both column and line sparklines to visual
ize the trends in the prices for each of the four stocks in the file. b. compute the daily change in the Dow Jones index
and apply a win/loss sparkline to visualize the daily up or down movement in the index. Dashboards 18. Create a useful dashboard for the data in the Excel
file President’s Inn Guest Database. Use the additional information stated in Problem 20 of Chapter 2: Room rates are the same for one or two guests;
135
however, additional guests must pay an additional $20 per person per day for meals. Guests staying for seven days or more receive a 10% discount. Modify the spreadsheet to calculate the number of days that each party stayed at the inn and the total revenue for the length of stay. Use appropriate charts and layouts and other visualization tools that help to convey the information. Explain why you chose the elements of the dashboard and how a manager might use them. 19. Create a useful dashboard for the data in the Excel
file Restaurant Sales. Use appropriate charts and layouts and other visualization tools that help to convey the information. Explain why you chose the elements of the dashboard and how a manager might use them. 20. Create a useful dashboard for the data in the Excel
file Store and Regional Sales Database. Use appropriate charts and layouts and other visualization tools that help to convey the information. Explain why you chose the elements of the dashboard and how a manager might use them. 21. Create a useful dashboard for the data in the Excel
file Employees. Use appropriate charts, layouts, and other visualization tools that help to convey the information. Explain why you chose the elements of the dashboard and how a manager might use them.
CASE: PERFORMANCE LAWN EQUIPMENT Part 1: PLE originally produced lawn mowers, but a significant portion of sales volume over recent years has come from the growing smalltractor market. As we noted in the case in Chapter 1, PLE sells their products worldwide, with sales regions including North America, South America, Europe, and the Pacific Rim. Three years ago, a new region was opened to serve China, where a booming market for small tractors has been established. PLE has always emphasized quality and considers the quality it builds into its products as its primary selling point. In the past two years, PLE has also emphasized the ease of use of their products. Before digging into the details of operations, Elizabeth Burke wants to gain an overview of PLE’s overall business performance and market position by examining the information provided in the database for this company. Specifically, she is asking you to construct appropriate charts for the data in the following worksheets in the Performance Lawn Equipment Database and summarize your conclusions from analysis of these charts.
a. Dealer Satisfaction b. EndUser Satisfaction c. Complaints d. Mower Unit Sales e. Tractor Unit Sales f. OnTime Delivery g. Defects After Delivery h. Response Time
Part 2: Propose a monthly dashboard of the most important business information that Ms. Burke can use on a routine basis as data are updated. Create one using the most recent data. Your dashboard should not consist of more than six to eight charts, which should fit comfortably on one screen. Write a formal report summarizing your results for both parts of this case.
APPENDIX
Additional Tools for Data Visualization
A3
In this appendix, we describe some additional charts that are only available in the Windows version of Excel 2016 and also illustrate the capabilities of Tableau, a powerful software package for data visualization.
Hierarchy Charts Hierarchical data are organized in a treelike structure. A simple example is the Purchase Orders database. At the top level we have suppliers; at the next level we have the items purchased; then the order quantity or cost per order, and so on. A treemap chart divides the chart area into rectangles that represent different levels and relative sizes of hierarchical data. Each rectangle is subdivided into smaller rectangles representing the next level in the hierarchy. Rectangles at the top level of the hierarchy are arranged with the largest in the upper left corner of the chart to the smallest in the lower right corner. Within a rectangle, the next level of the hierarchy is also arranged with rectangles from the upper left to the lower right. EXAMPLE A3.1
A Treemap for Purchase Order Data
For the Purchase Orders database, we used a PivotTable to create a new worksheet showing the items purchased from each supplier and the total cost of the orders. Highlight the data and select Insert Hierarchy Chart from the Charts group on the Insert tab. Figure A3.1 shows the
result. We see that the largest order costs come from Hulkey Fasteners, followed by Durrable Products, Steelpin Inc., and so on. Within each colored area you can also see the relative sizes of the orders from each supplier.
An alternative to a treemap chart is a sunburst chart. In this chart, the hierarchy is represented by a series of circles, with the highest level of the hierarchy in the center and lower levels of the hierarchy as rings displayed outside the center. The lowest level of the hierarchy is the outside ring. Figure A3.2 shows a sunburst chart for the Purchase Orders data in the previous example.
Waterfall Charts A waterfall chart shows a running total as values are added or subtracted. A common application of waterfall charts is to show the impact of positive or negative cash flows on net income or profit. 136
▲▲ Figure
Appendix A3 Additional Tools for Data Visualization
A3.1
Purchase Orders Treemap
▶▶ Figure
A3.2
Sunburst Chart
137
138
Appendix A3 Additional Tools for Data Visualization
EXAMPLE A3.2
Creating a Waterfall Chart
Figure A3.3 shows the calculation of net income as sales less cost of goods sold, administrative expenses, selling expenses, depreciation expenses, interest expense, and taxes (these data are found in a slightly different format in the Excel file Net Income Models, which we will use in Chapter 11). Simply highlight the range A3:B10, choose
Insert Waterfall or Stock Chart from the Charts group on the Insert tab, and select the waterfall chart. In the default chart, double click on the net income data point. In the Format Data Point window, check the box Set as total. This positions the net income bar on the xaxis. The chart shows how each component adds to or decreases the net income.
PivotCharts Microsoft Excel for Windows provides a simple oneclick way of creating charts to visualize data in PivotTables called PivotCharts. To display a PivotChart for a PivotTable, first click inside the PivotTable. From the Analyze tab, click on PivotChart. Excel will display an Insert Chart dialog that allows you to choose the type of chart you wish to display. EXAMPLE A3.3
A PivotChart for Sales Transactions Data
Figure 2.34 showed a PivotTable for revenues by region and product for the Sales Transactions database. To display a column PivotChart, choose Clustered Column from the
PivotChart menu. Figure A3.4 shows the chart generated by Excel. By clicking on the dropdown filter buttons, you can easily change the display.
PivotCharts make it easy to create dashboards. However, you should consider what message you are trying to convey in choosing the design of PivotCharts. In Figure A3.4, for instance, it is easier to compare product sales within a region, while in Figure A3.5, which was created by reversing the order of region and product in the Rows area, it is easier to compare regional sales for a given product. PivotCharts are also highly effective in oral presentations since the Filter buttons allow you to drill down into the data to respond to questions from the audience.
▲▲ Figure
A3.3
Waterfall Chart
Appendix A3 Additional Tools for Data Visualization
▶▶ Figure
139
A3.4
PivotChart for Revenues by Region and Product
▶▶ Figure
A3.5
Alternate PivotChart for Revenues by Region and Product
Tableau Tableau (www.tableau.com) is professional business intelligence software that facilitates data visualization using intuitive draganddrop capabilities. It can be used for spreadsheets as well as other types of database formats such as geographic data and has the power to handle big data easily. It also applies best practices in data visualization and facilitates the development of dashboards, which can easily be published and shared on the web and on mobile devices. We briefly illustrate some of its basic capabilities. Figure A3.6 shows the window for the Purchase Orders database after it has been loaded into Tableau. Tableau automatically partitions the data in the left pane as “dimensions,” namely, the columns that correspond to names or dates, and “measures,” those that
140
Appendix A3 Additional Tools for Data Visualization
correspond to numerical data in the database. Similar to PivotTables, you simply drag these into the areas for Columns, Rows, Filters, and Pages. In Figure A3.6, we first drag Item Description into the Columns area and Supplier into the Rows area; then drag Cost per order inside of the rectangular area formed by the rows and columns. Note that this displays SUM(Cost per order) as shown in the green oval. From a dropdown menu, you can change this to average, count, minimum, maximum, and other measures, similar to changing the field settings in a PivotTable. The Show Me window provides oneclick visualizations for a wide variety of charts, including common bar, column, and pie charts, along with more advanced charts such as treemaps and heat maps. Tableau intelligently shows only those charts that apply to the data and hides the others (grayed out). Figure A3.7 shows a stacked bar chart, and Figure A3.8 shows a bubble chart view of the data. Tableau provides free licenses for classroom use and a host of training materials and instructional videos. We encourage you to visit www.tableau.com/products/desktop to see more of the capabilities and features of Tableau. ▶▶ Figure
A3.6
Tableau Window
▶▶ Figure
A3.7
Tableau Stacked Bar Chart
Appendix A3 Additional Tools for Data Visualization
▶▶ Figure
141
A3.8
Tableau Bubble Chart
PROBLEMS AND EXERCISES Hierarchy Charts 1. Create a treemap for the regions, items, and units sold
in the Excel file Store and Regional Sales Database. 2. Create a sunburst chart for the regions, items, and
units sold in the Excel file Store and Regional Sales Database. 3. Compute the total expenses for each month in the
Budget Forecasting Excel file. Then develop waterfall charts for each month to visualize the data. PivotCharts 4. Use PivotTables to find the number of loans by dif
ferent purposes, marital status, and credit risk in the Excel file Credit Risk Data. Illustrate the results in a PivotChart. 5. Use PivotCharts to create a useful dashboard for the
data in the Excel file President’s Inn Guest Database. Explain why you chose the elements of the dashboard and how a manager might use them.
6. Use PivotCharts to create a useful dashboard for the
data in the Excel file Restaurant Sales. Explain why you chose the elements of the dashboard and how a manager might use them. 7. Use PivotCharts to create a useful dashboard for
the data in the Excel file Store and Regional Sales. Explain why you chose the elements of the dashboard and how a manager might use them. 8. Use PivotCharts to create a useful dashboard for the
data in the Excel file Corporate Default Database. Explain why you chose the elements of the dashboard and how a manager might use them. 9. Use PivotCharts to create a useful dashboard for the
data in the Excel file Peoples Choice Bank. Explain why you chose the elements of the dashboard and how a manager might use them.
This page intentionally left blank
CHAPTER
4
Descriptive Statistics
Nataliiap/Shutterstock
LEARNING OBJECTIVES ■■
■■
■■
■■
■■
■■
After studying this chapter, you will be able to:
Define a metric and explain the concepts of measurement and measures. Explain the difference between a discrete metric and continuous metric, and provide examples of each. Describe the four groups of data classification: categorical, ordinal, interval, and ratio, and provide examples of each. Explain the science of statistics and define the term statistic. Construct a frequency distribution for categorical, numerical, and grouped data. Construct a relative frequency distribution and histogram.
■■ ■■ ■■ ■■
■■ ■■ ■■
■■
■■
Compute cumulative relative frequencies. Find percentiles and quartiles for a data set. Construct a crosstabulation (contingency table). Explain the difference between a population and a sample. Understand statistical notation. List different measures of location. Compute the mean, median, mode, and midrange of a set of data. Use measures of location to make practical business decisions. List different measures of dispersion.
144
Chapter 4 Descriptive Statistics
Compute the range, interquartile range, variance, and standard deviation of a set of data. ■■ Explain Chebyshev’s theorem. ■■ State the empirical rules and apply them to practical data. ■■ Compute a standardized value (zscore) for observations in a data set. ■■ Define and compute the coefficient of variation. ■■ Explain the nature of skewness and kurtosis in a distribution. ■■ Interpret the coefficients of skewness and kurtosis. ■■ Use the Excel Descriptive Statistics tool to summarize data. ■■ Calculate the mean, variance, and standard deviation for grouped data. ■■
■■ ■■
■■
■■ ■■ ■■ ■■ ■■
■■
Calculate a proportion. Use PivotTables to compute the mean, variance, and standard deviation of summarized data. Explain the importance of understanding relationships between two variables. Explain the difference between covariance and correlation. Calculate measures of covariance and correlation. Use the Excel Correlation tool. Identify outliers in data. State the principles of statistical thinking. Interpret variation in data from a logical and practical perspective. Explain the nature of variation in sample data.
Statistics, as defined by David Hand, past president of the Royal Statistical Society in the UK, is both the science of uncertainty and the technology of extracting information from data.1 Statistics involves collecting, organizing, analyzing, interpreting, and presenting data. A statistic is a summary measure of data. You are undoubtedly familiar with the concept of statistics in daily life as reported in the media: baseball batting averages, airline ontime arrival performance, and economic statistics such as the Consumer Price Index are just a few examples. Statistics helps to understand and quantify uncertainty in data and to incorporate uncertainty for predicting the future. You have undoubtedly been using statistics informally for many years. Numerical measures, such as your GPA, average incomes, and housing prices, are examples of statistics. While we see averages all the time in sports, finance, and marketing, these are only the tip of the iceberg. Statistics is much more than just computing averages. Statistics provides the means of gaining insight—both numerically and visually—into large quantities of data, understanding uncertainty and risk, and drawing conclusions from sample data that come from very large populations. For example, marketing analysts employ statistics extensively to analyze survey data to understand brand loyalty and satisfaction with goods and services, to segment customers into groups for targeted ads, and to identify factors that drive consumer demand; finance personnel use statistics to evaluate stock and mutual fund performance in order to identify good investment opportunities and to evaluate changes in foreign currency rates; and operations managers use statistics to gauge production and quality performance to determine
1 David Hand, “Statistics: An Overview,” in Miodrag Lovric, Ed., International Encyclopedia of Statistical Science, Springer Major Reference; http://www.springer.com/statistics/book/9783642048975, p. 1504.
Chapter 4 Descriptive Statistics
145
process and design improvements. You will find yourself routinely using many of the statistical concepts in this chapter in your daily work. Statistical methods are essential to business analytics and are used throughout this book. Microsoft Excel supports statistical analysis in two ways: 1. with statistical functions that are entered in worksheet cells directly or embedded in formulas 2. with the Excel Analysis Toolpak addin to perform more complex statistical computations. While previous versions of Excel for Mac did not support the Analysis Toolpak, Excel 2016 now includes it. We use both statistical functions and the Analysis Toolpak in many examples. Descriptive statistics refers to methods of describing and summarizing data using tabular, visual, and quantitative techniques. In this chapter, we focus on both tabular and visual methods, as well as quantitative measures for statistical analysis of data. We begin with a discussion of different types of metrics and ways to classify data.
ANALYTICS IN PRACTICE: Applications of Statistics in Health Care2 The science of statistics is vitally important to health care decision makers. Descriptive statistics summarize the utility, efficacy, and costs of medical goods and services. Increasingly, health care organizations employ statistical analysis to measure their performance outcomes. Some examples include the following: ■■
■■
2
Descriptive statistics summarize the utility, efficacy, and costs of medical goods and services. For example, government health and human service agencies gauge the overall health and wellbeing of populations with statistical information. Hospitals and other large provider service organizations implement datadriven, continuous quality improvement programs to maximize efficiency. Statistics are important to health care companies in measuring performance success or failure. By establishing benchmarks, or standards of service excellence, quality improvement managers can measure future outcomes. Analysts map the overall growth and viability of a health care company by using statistical data gathered over time.
■■
■■
■■
■■
Researchers gather data on human population samples. The health care industry benefits from knowing consumer market characteristics such as age, sex, race, income, and disabilities. These demographic statistics can predict the types of services that people are using and the level of care that is affordable to them. Statistical information is invaluable in determining what combination of goods and services to produce, which resources to allocate in producing them, and to which populations to offer them. Public and private health care administrators, charged with providing continuums of care to diverse populations, compare existing services to community needs. Statistical analysis is a critical component in a needs assessment. Statistics is equally important to pharmaceutical and technology companies in developing product lines that meet the needs of the populations they serve. Innovative medicine begins and, sometimes, ends with statistical analysis. Data are collected and carefully reported in clinical trials of new technologies and treatments to weigh products’ benefits against their risks.
Adapted from Rae Casto, “Why Are Statistics Important in the Health Care Field?” https://www.livestrong.com/article/ 186334whyarestatisticsimportantinthehealthcarefield/.
146
Chapter 4 Descriptive Statistics
Metrics and Data Classification A metric is a unit of measurement that provides a way to objectively quantify performance. For example, senior managers might assess overall business performance using such metrics as net profit, return on investment, market share, and customer satisfaction. A plant manager might monitor such metrics as the proportion of defective parts produced or the number of inventory turns each month. For a Webbased retailer, some useful metrics are the percentage of orders filled accurately and the time taken to fill a customer’s order. Measurement is the act of obtaining data associated with a metric. Measures are numerical values associated with a metric. Metrics can be either discrete or continuous. A discrete metric is one that is derived from counting something. For example, a delivery is either on time or not; an order is complete or incomplete; or an invoice can have one, two, three, or any number of errors. Some discrete metrics associated with these examples would be the number of ontime deliveries, the number of incomplete orders each day, and the number of errors per invoice. Continuous metrics are based on a continuous scale of measurement. Any metrics involving dollars, length, time, volume, or weight, for example, are continuous. Another classification of data is by the type of measurement scale. Data may be classified into four groups: 1. Categorical (nominal) data, which are sorted into categories according to specified characteristics. For example, a firm’s customers might be classified by their geographical region (e.g., North America, South America, Europe, and Pacific); employees might be classified as managers, supervisors, and associates. The categories bear no quantitative relationship to one another, but we usually assign an arbitrary number to each category to ease the process of managing the data and computing statistics. Categorical data are usually counted or expressed as proportions or percentages. 2. Ordinal data, which can be ordered or ranked according to some relationship to one another. College football or basketball rankings are ordinal; a higher ranking signifies a stronger team but does not specify any numerical measure of strength. Ordinal data are more meaningful than categorical data because data can be compared. A common example in business is data from survey scales—for example, rating a service as poor, average, good, very good, or excellent. Such data are categorical but also have a natural order (excellent is better than very good) and, consequently, are ordinal. However, ordinal data have no fixed units of measurement, so we cannot make meaningful numerical statements about differences between categories. Thus, we cannot say that the difference between excellent and very good is the same as between good and average, for example. Similarly, a team ranked number 1 may be far superior to the number 2 team, whereas there may be little difference between teams ranked 9th and 10th. 3. Interval data, which are ordinal but have constant differences between observations and have arbitrary zero points. Common examples are time and temperature. Time is relative to global location, and calendars have arbitrary starting dates (compare, for example, the standard Gregorian calendar with the Chinese calendar). Both the Fahrenheit and Celsius scales represent a specified measure of distance—degrees—but have arbitrary zero points. Thus we cannot compute meaningful ratios; for example, we cannot say that 50 degrees is twice as hot as 25 degrees. However, we can compare differences. Another example is SAT or GMAT scores. The scores can be used to rank students, but only differences
Chapter 4 Descriptive Statistics
147
between scores provide information on how much better one student performed over another; ratios make little sense. In contrast to ordinal data, interval data allow meaningful comparison of ranges, averages, and other statistics. In business, data from survey scales, while technically ordinal, are often treated as interval data when numerical scales are associated with categories (for instance, 1 = poor, 2 = average, 3 = good, 4 = very good, 5 = excellent). Strictly speaking, this is not correct because the “distance” between categories may not be perceived as the same (respondents might perceive a larger gap between poor and average than between good and very good, for example). Nevertheless, many users of survey data treat them as interval when analyzing the data, particularly when only a numerical scale is used without descriptive labels. 4. Ratio data, which are continuous and have a natural zero point. Most business and economic data, such as dollars and time, fall into this category. For example, the measure dollars has an absolute zero. Ratios of dollar figures are meaningful. For example, knowing that the Seattle region sold $12 million in March whereas the Tampa region sold $6 million means that Seattle sold twice as much as Tampa. This classification is hierarchical in that each level includes all the information content of the one preceding it. For example, ordinal data are also categorical, and ratio information can be converted to any of the other types of data. Interval information can be converted to ordinal or categorical data but cannot be converted to ratio data without the knowledge of the absolute zero point. Thus, a ratio scale is the strongest form of measurement. EXAMPL E 4.1
Classifying Data Elements in a Purchasing Database3 Item description—categorical Item cost—ratio Quantity—ratio Cost per order—ratio A/P terms—ratio Order date—interval Arrival date—interval
Figure 4.1 shows a portion of a data set containing all items that an aircraft component manufacturing company has purchased over the past three months. The data provide the supplier; order number; item number, description, and cost; quantity ordered; cost per order; the suppliers’ accounts payable (A/P) terms; and the order and arrival dates. We may classify each of these types of data as follows:
■■ ■■ ■■ ■■ ■■ ■■ ■■
■■ Supplier—categorical ■■ Order number—ordinal ■■ Item number—categorical
We might use these data to evaluate the average speed of delivery and rank the suppliers (thus creating ordinal data) by this metric.
▲▲ Figure
4.1
Portion of Purchase Orders Database 3 Based on Kenneth C. Laudon and Jane P. Laudon, Essentials of Management Information Systems. 9th ed. (Upper Saddle River, NJ: Prentice Hall, 2011).
148
Chapter 4 Descriptive Statistics
CHECK YOUR UNDERSTANDING 1. Explain the science of statistics. 2. What is a metric? How does it differ from a measure? 3. Explain the difference between a discrete and a continuous metric. 4. Describe the four types of measurement scales and give an example of each.
Frequency Distributions and Histograms A frequency distribution is a table that shows the number of observations in each of several nonoverlapping groups. A graphical depiction of a frequency distribution in the form of a column chart is called a histogram. Frequency distributions and histograms summarize basic characteristics of data, such as where the data are centered and how broadly data are dispersed. This is usually the first step in using descriptive statistics. In this section, we discuss how to create them for both categorical and numerical data.
Frequency Distributions for Categorical Data Categorical variables naturally define the groups in a frequency distribution. For example, in the Purchase Orders database (see Figure 4.1), orders were placed for the following items: Airframe fasteners Boltnut package Control panel Door decal Electrical connector Gasket Hatch decal
Machined valve Oring Panel decal Pressure gauge Shielded cable/ft. Side panel
To construct a frequency distribution, we need only count the number of observations that appear in each category. This can be done using the Excel COUNTIF function.
EXAMPLE 4.2 Constructing a Frequency Distribution for Items in
the Purchase Orders Database First, list the item names in a column on the spreadsheet. We used column A, starting in cell A100, below the existing data array. It is important to use the exact names as used in the data file. To count the number of orders placed for each item, use the function = COUNTIF($D$4:$D$97, cell_reference), where cell_reference is the cell containing the item name, in this case, cell A101. This is shown in Figure 4.2.
The resulting frequency distribution for the items is shown in Figure 4.3. Thus, the company placed 14 orders for airframe fasteners and 11 orders for the boltnut package. We may also construct a column chart to visualize these frequencies, as shown in Figure 4.4. We might wish to sort these using Pareto analysis to gain more insight into the order frequency.
▶▶ Figure
Chapter 4 Descriptive Statistics
149
4.2
Using the COUNTIF Function to Construct a Frequency Distribution
▶▶ Figure
4.3
Frequency Distribution for Items Purchased
▶▶ Figure
4.4
Column Chart for Frequency Distribution of Items Purchased
Relative Frequency Distributions We may express the frequencies as a fraction, or proportion, of the total; this is called the relative frequency. If a data set has n observations, the relative frequency of category i is computed as
Relative Frequency of Category i =
Frequency of Category i (4.1) n
We often multiply the relative frequencies by 100 to express them as percentages. A relative frequency distribution is a tabular summary of the relative frequencies of all categories.
150
Chapter 4 Descriptive Statistics
▶▶ Figure
4.5
Relative Frequency Distribution for Items Purchased
EXAMPLE 4.3 Constructing a Relative Frequency Distribution
for Items in the Purchase Orders Database The calculations for relative frequencies are simple. First, sum the frequencies to find the total number (note that the sum of the frequencies must be the same as the total number of observations, n). Then divide the frequency of each category by this value. Figure 4.5 shows the relative frequency distribution for the purchase order items. The formula in cell C101, for example, is = B101/$B$114. You then
copy this formula down the column to compute the other relative frequencies. Note that the sum of the relative frequencies must equal 1.0. A pie chart of the frequencies is sometimes used to show these proportions visually, although it is more appealing for a smaller number of categories. For a large number of categories, a column or bar chart would work better, as we noted in Chapter 3.
Frequency Distributions for Numerical Data For numerical data that consist of a small number of values, we may construct a frequency distribution similar to the way we did for categorical data; that is, we simply use COUNTIF to count the frequencies of each value. EXAMPLE 4.4
Frequency and Relative Frequency Distribution for A/P Terms
In the Purchase Orders data, the A/P terms are all whole numbers: 15, 25, 30, and 45. A frequency and relative frequency distribution for these data are shown in Figure 4.6.
A bar chart showing the proportions, or relative frequencies, in Figure 4.7, clearly shows that the majority of orders had accounts payable terms of 30 months.
Excel Histogram Tool Frequency distributions and histograms can be created using the Analysis Toolpak in Excel. To do this, click the Data Analysis tools button in the Analysis group under the Data tab in the Excel menu bar and select Histogram from the list. In the dialog box (see Figure 4.8), specify the Input Range corresponding to the data. If you include the column header, then also check the Labels box so Excel knows that the range contains a label. The ▶▶ Figure
4.6
Frequency and Relative Frequency Distribution for A/P Terms
Chapter 4 Descriptive Statistics
▶▶ Figure
151
4.7
Bar Chart of Relative Frequencies of A/P Terms
Bin Range defines the groups (Excel calls these “bins”) used for the frequency distribution. If you do not specify a Bin Range, Excel will automatically determine bin values for the frequency distribution and histogram, which often results in a rather poor choice. If you have discrete values, set up a column of these values in your spreadsheet for the bin range and specify this range in the Bin Range field. We describe how to handle continuous data shortly. Check the Chart Output box to display a histogram in addition to the frequency distribution. You may also sort the values as a Pareto chart and display the cumulative frequencies by checking the additional boxes. EXAMPL E 4.5
Using the Histogram Tool
We will create a frequency distribution and histogram for the A/P Terms variable in the Purchase Orders database. Figure 4.9 shows the completed Histogram dialog. The input range includes the column header as well as the data in column H. We define the bin range below and enter it in cells H99:H103 (including the header “Months”):
If you check the Labels box, it is important that both the Input Range and the Bin Range have labels included in the first row. Figure 4.10 shows the results from this tool.
Months 15 25 30 45
Grouped Frequency Distributions For numerical data that have many different discrete values with little repetition or are continuous, we usually group the data into “bins.” A grouped frequency distribution requires that we specify 1. the number of groups, 2. the width of each group, and 3. the upper and lower limits of each group.
152
Chapter 4 Descriptive Statistics
▶▶ Figure
4.8
Histogram Tool Dialog
▶▶ Figure
4.9
Histogram Dialog for A/P Terms Data
▶▶ Figure
4.10
Excel Frequency Distribution and Histogram for A/P Terms
It is important to remember that the groups may not overlap, so that each value is counted in exactly one group. You should define the groups after examining the range of the data. Generally, you should choose between 5 and 15 groups, and the range of each should be equal. The more data you have, the more groups you should generally use. Note that with fewer groups, the group widths will be wider. Wider group widths provide a “coarse” histogram. Sometimes you need to experiment to find the best number of groups to provide a useful visualization of the data. Choose the lower limit of the first group (LL) as a whole number smaller than the minimum data value and the upper limit of the last group (UL) as a whole number
Chapter 4 Descriptive Statistics
153
larger than the maximum data value. Generally, it makes sense to choose nice, round whole numbers. Then you may calculate the group width as
Group Width =
UL  LL (4.2) Number of Groups
EXAMPL E 4.6 Constructing a Frequency Distribution and Histogram
for Cost per Order In this example, we apply the Excel Histogram tool to the cost per order data in column G of the Purchase Orders database. The data range from a minimum of $68.75 to a maximum of $127,500. You can find this either by using the MIN and MAX functions or simply by sorting the data. To ensure that all the data will be included in some group, it makes sense to set the lower limit of the first group to $0 and the upper limit of the last group to $130,000. Thus, if we select five groups, using equation (4.2), the width of each group is 1$130,000  02 >5 = $26,000; if we choose ten groups, the width is 1$130,000  02 >10 = $13,000. We select five groups. Doing so, the bin range is specified as
Upper Group Limit $ 0.00 $ 26,000.00 $ 52,000.00 $ 78,000.00 $104,000.00 $130,000.00
▶▶ Figure
4.11
Frequency Distribution and Histogram for Cost per Order (Five Groups)
This means that the first group includes all values less than or equal to $0; the second group includes all values greater than $0 but less than or equal to $26,000, and so on. Note that the groups do not overlap because the lower limit of one group is strictly greater than the upper limit of the previous group. We suggest using the header “Upper Group Limit” for the bin range to make this clear. In the spreadsheet, this bin range is entered in cells G99:G105. The Input Range in the Histogram dialog is G4:G97. Figure 4.11 shows the results. These results show that the vast majority of orders were for $26,000 or less and fall rapidly beyond this value. Selecting a larger number of groups might help to better understand the nature of the data. Figure 4.12 shows results using ten groups. This shows that there were a higher percentage of orders for $13,000 or less than for between $13,000 and $26,000.
154
Chapter 4 Descriptive Statistics
▶▶ Figure
4.12
Frequency Distribution and Histogram for Cost per Order (Ten Groups)
One limitation of the Excel Histogram tool is that the frequency distribution and h istogram are not linked to the data; thus, if you change any of the data, you must repeat the entire procedure to construct a new frequency distribution and histogram.
Cumulative Relative Frequency Distributions For numerical data, we may also compute the relative frequency of observations in each group. By summing all the relative frequencies at or below each upper limit, we obtain the cumulative relative frequency. The cumulative relative frequency represents the proportion of the total number of observations that fall at or below the upper limit of each group. A tabular summary of cumulative relative frequencies is called a cumulative relative frequency distribution.
EXAMPLE 4.7
Computing Cumulative Relative Frequencies
Figure 4.13 shows the relative frequency and cumulative relative frequency distributions for the cost per order data in the Purchase Orders database using ten groups. The relative frequencies are computed using the same approach as in Example 4.3—namely, by dividing the frequency by the total number of observations (94). In column D, we set the cumulative relative frequency of the first group equal to its relative frequency. Then we add the relative frequency
of the next group to the cumulative relative frequency. For example, the cumulative relative frequency in cell D3 is computed as = D2 +C3 = 0.000 +0.4468 = 0.4468; the cumulative relative frequency in cell D4 is computed as = D3 +C4 = 0.4468 +0.2766 = 0.7234, and so on. Because relative frequencies must be between 0 and 1 and must add up to 1, the cumulative frequency for the last group must equal 1.
Figure 4.14 shows a chart for the cumulative relative frequency, which is called an ogive. From this chart, you can easily estimate the proportion of observations that fall below a certain value. For example, you can see that slightly more than 70% of the data fall at or below $26,000, about 90% of the data fall at or below $78,000, and so on. Note that cumulative frequencies can also be displayed using the Histogram tool.
Chapter 4 Descriptive Statistics
▶▶ Figure
155
4.13
Cumulative Relative Frequency Distribution for Cost per Order Data
▶▶ Figure
4.14
Ogive for Cost per Order
Constructing Frequency Distributions Using PivotTables PivotTables make it quite easy to construct frequency distributions. For example, in the Purchase Orders data, we can simply build a PivotTable to find a count of the number of orders for each item, resulting in the same summary we saw in Figure 4.5. With Excel for Windows, a PivotChart (see Appendix A3 in Chapter 3) will display the histogram. For continuous numerical data, we can also use PivotTables to construct a grouped frequency distribution. EXAMPL E 4.8
Constructing a Grouped Frequency Distribution Using PivotTables
Using the Purchase Orders database, create a PivotTable as shown in Figure 4.15. Note that this simply shows the frequencies of the cost per order values, each of which is unique. It doesn’t matter what field is used in the Sum Values area; just ensure that the field settings specify “count.” Next, click on any value in the Row Labels column,
and from the Analyze tab for PivotTable Tools, select Group Field. Edit the dialog to start at 0 and end at 130000, and use 26000 as the group range, as shown in Figure 4.16. The result is shown in Figure 4.17; this groups the data in the same bins as we saw in Figure 4.11.
156
Chapter 4 Descriptive Statistics
▶▶ Figure
4.15
Count of Cost per Order PivotTable
▶▶ Figure
4.16
Group Field Dialog
▶▶ Figure
4.17
Grouped Frequency Distribution in a PivotTable
CHECK YOUR UNDERSTANDING 1. What is a frequency distribution? 2. Explain how to calculate a relative frequency. 3. State how to use Excel for constructing frequency distributions for categorical and numerical data. 4. State the features and limitations of the Excel Histogram tool. 5. What does an ogive display? 6. Use a PivotTable to construct the frequency distribution in Figure 4.12.
Chapter 4 Descriptive Statistics
157
Percentiles and Quartiles Data are often expressed as percentiles and quartiles. You are no doubt familiar with percentiles from standardized tests used for college or graduate school entrance examinations (SAT, ACT, GMAT, GRE, etc.). Percentiles specify the percent of other test takers who scored at or below the score of a particular individual. Generally speaking, the kth percentile is a value at or below which at least k percent of the observations lie. However, the way by which percentiles are calculated is not standardized. The most common way to compute the kth percentile is to order the data values from smallest to largest and calculate the rank of the kth percentile using the formula nk + 0.5 (4.3) 100 where n is the number of observations. Round this to the nearest integer, and take the value corresponding to this rank as the kth percentile.
EXAMPL E 4.9
Computing Percentiles
In the Purchase Orders data, we have n = 94 observations. The rank of the 90th percentile (k = 90) for the cost per order data is computed as 94(90)>100 + 0.5 = 85.1, or,
rounded, 85. The 85th ordered value is $74,375 and is the 90th percentile. This means that 90% of the costs per order are less than or equal to $74,375, and 10% are higher.
Statistical software use different methods that often involve interpolating between ranks instead of rounding, thus producing different results. The Excel function PERCENTILE.INC(array, k) computes the 100 * kth percentile of data in the range specified in the array field, where k is in the range 0 to 1, inclusive. EXAMPL E 4.10 Computing Percentiles in Excel To find the 90th percentile for the cost per order data in the Purchase Orders data, use the Excel function = PERCENTILE.INC(G4:G97,0.9). This calculates the 90th
percentile as $73,737.50, which is different from using formula (4.3).
Excel also has a tool for sorting data from high to low and computing percentiles associated with each value. Select Rank and Percentile from the Data Analysis menu in the Data tab and specify the range of the data in the dialog. Be sure to check the Labels in First Row box if your range includes a header in the spreadsheet. EXAMPL E 4.11 Excel Rank and Percentile Tool A portion of the results from the Rank and Percentile tool for the cost per order data is shown in Figure 4.18. You can see
that the Excel value of the 90th percentile that we computed in Example 4.9 as $74,375 is the 90.3rd percentile value.
Quartiles break the data into four parts. The 25th percentile is called the first quartile, Q1; the 50th percentile is called the second quartile, Q2; the 75th percentile is called the third quartile, Q3; and the 100th percentile is the fourth quartile, Q4. Onefourth of the data fall below the first quartile, onehalf are below the second quartile, and threefourths are below the third quartile. We may compute quartiles using the Excel function Q UARTILE.INC(array, quart), where array specifies the range of the data and quart is a whole number between 1 and 4, designating the desired quartile.
158
Chapter 4 Descriptive Statistics
▶▶ Figure
4.18
Portion of Rank and Percentile Tool Results
EXAMPLE 4.12 Computing Quartiles in Excel For the cost per order data in the Purchase Orders database, we may use the Excel function = QUARTILE.INC(G4:G97,k), where k ranges from 1 to 4, to compute the quartiles. The results are as follows:
k ∙ 1
First quartile
k ∙ 2
Second quartile
$15,656.25
k ∙ 3
Third quartile
$27,593.75
k ∙ 4
Fourth quartile
We may conclude that 25% of the order costs fall at or below $6,757.81; 50% fall at or below $15,656.25; 75% fall at or below $27,593.75, and 100% fall at or below the maximum value of $127,500.
$6,757.81
$127,500.00
We can extend these ideas to other divisions of the data. For example, deciles divide the data into ten sets: the 10th percentile, 20th percentile, and so on. All these types of measures are called data profiles, or fractiles. CHECK YOUR UNDERSTANDING 1. Explain how to interpret the 75th percentile. 2. How do quartiles relate to percentiles? 3. What Excel functions can you use to find percentiles and quartiles?
CrossTabulations One of the most basic statistical tools used to summarize categorical data and examine the relationship between two categorical variables is crosstabulation. A crosstabulation is a tabular method that displays the number of observations in a data set for different subcategories of two categorical variables. A crosstabulation table is often called a contingency table. The subcategories of the variables must be mutually exclusive and exhaustive, meaning that each observation can be classified into only one subcategory, and, taken together over all subcategories, they must constitute the complete data set. Crosstabulations are commonly used in marketing research to provide insight into characteristics of different market segments using categorical variables such as gender, educational level, and marital status.
Chapter 4 Descriptive Statistics
EXAMPL E 4.13
Constructing a CrossTabulation
Let us examine the Sales Transactions database, a portion of which is shown in Figure 4.19. Suppose we wish to identify the number of books and DVDs ordered by region. A crosstabulation will have rows corresponding to the different regions and columns corresponding to the products. Within the table we list the count of the number in each pair of categories. A crosstabulation of these data is shown in Table 4.1. PivotTables make it easy to construct crosstabulations. Visualizing the data as a chart is a good way of communicating the results. Figure 4.20 shows the differences between product and regional sales. Expressing the results as percentages of a row or column makes it easier to interpret differences between
▶▶ Figure
regions or products, particularly as the totals for each category differ. Table 4.2 shows the percentage of book and DVD sales within each region; this is computed by dividing the counts by the row totals and multiplying by 100 (in Excel, simply divide the count by the total and format the result as a percentage by clicking the % button in the Number group within the Home tab in the ribbon). In Chapter 2, we saw how to do this easily in a PivotTable. For example, we see that although more books and DVDs are sold in the West region than in the North, the relative percentages of each product are similar, particularly when compared to the East and South regions.
4.19
Portion of Sales Transactions Database
▶▶
Table 4.1
CrossTabulation of Sales Transaction Data
▶▶
Table 4.2
Percentage Sales of Products Within Each Region
159
Region
Book
DVD
Total
East
56
42
98
North
43
42
85
South
62
37
99
West
100
90
190
Total
261
211
472
Region
Book
DVD
Total
East
57.1%
42.9%
100.0%
North
50.6%
49.4%
100.0%
South
62.6%
37.4%
100.0%
West
52.6%
47.4%
100.0%
160
Chapter 4 Descriptive Statistics
▶▶ Figure
4.20
Chart of Regional Sales by Product
CHECK YOUR UNDERSTANDING 1. What is a crosstabulation? 2. Use a PivotTable to construct the crosstabulation shown in Table 4.1.
Descriptive Statistical Measures In this section, we introduce numerical measures that provide an effective and efficient way of obtaining meaningful information from data. Before discussing these measures, however, we need to understand the differences between populations and samples.
Populations and Samples A population consists of all items of interest for a particular decision or investigation—for example, all individuals in the United States who do not own cell phones, all subscribers to Netflix, or all stockholders of Google. A company like Netflix keeps extensive records on its customers, making it easy to retrieve data about the entire population of customers. However, it would probably be impossible to identify all individuals who do not own cell phones. A sample is a subset of a population. For example, a list of individuals who rented a comedy from Netflix in the past year would be a sample from the population of all customers. Whether this sample is representative of the population of customers—which depends on how the sample data are intended to be used—may be debatable; nevertheless, it is a sample. Most populations, even if they are finite, are generally too large to deal with effectively or practically. For instance, it would be impractical as well as too expensive to survey the entire population of TV viewers in the United States. Sampling is also clearly necessary when data must be obtained from destructive testing or from a continuous production process. Thus, the purpose of sampling is to obtain sufficient information to draw a valid inference about a population. Market researchers, for example, use sampling to gauge consumer perceptions on new or existing goods and services; auditors use sampling to verify the accuracy of financial statements; and qualitycontrol analysts sample production output to verify quality levels and identify opportunities for improvement. Most data with which businesses deal are samples. For instance, the Purchase Orders and Sales Transactions databases that we used in previous chapters represent samples because the purchase order data include only orders placed within a threemonth time
161
Chapter 4 Descriptive Statistics
period, and the sales transactions represent orders placed on only one day, July 14. Therefore, unless it is obvious or noted otherwise, we will assume that any data set is a sample.
Statistical Notation We typically label the elements of a data set using subscripted variables, x1, x2, and so on. In general, xi represents the ith observation. It is a common practice in statistics to use Greek letters, such as m (mu), s (sigma), and p (pi), to represent population measures and italic letters such as x (xbar), s, and p to represent sample statistics. We will use N to represent the number of items in a population and n to represent the number of observations in a sample. Statistical formulas often contain a summation operator, Σ (Greek capital sigma), which means that the terms that follow it are added together. Thus, a xi = x1 + x2 + g + xn. Understanding these conventions and mathematical n
i=1
notation will help you to interpret and apply statistical formulas.
Measures of Location: Mean, Median, Mode, and Midrange Measures of location provide estimates of a single value that in some fashion represents the “centering” of a set of data. The most common is the average. We all use averages routinely in our lives, for example, to measure student accomplishment in college (e.g., grade point average), to measure the performance of athletes (e.g., batting average), and to measure performance in business (e.g., average delivery time). Arithmetic Mean The average is formally called the arithmetic mean (or simply the mean), which is the sum of the observations divided by the number of observations. Mathematically, the mean of a population is denoted by the Greek letter m, and the mean of a sample is denoted by x. If a population consists of N observations x1, x2,c, xN, the population mean, m, is calculated as a xi N
m =
i=1
N
(4.4)
The mean of a sample of n observations, x1, x2,c, xn, denoted by x, is calculated as a xi n
x =
i=1
n
(4.5)
Note that the calculations for the mean are the same whether we are dealing with a population or a sample; only the notation differs. We may also calculate the mean in Excel using the function AVERAGE(data range). One property of the mean is that the sum of the deviations of each observation from the mean is zero:
a 1xi  x2 = 0 i
(4.6)
This simply means that the sum of the deviations above the mean is the same as the sum of the deviations below the mean; essentially, the mean “balances” the values on either side of
162
Chapter 4 Descriptive Statistics
it. However, it does not suggest that half the data lie above or below the mean—a common misconception among those who don’t understand statistics. In addition, the mean is unique for every set of data and is meaningful for both interval and ratio data. However, it can be affected by outliers—observations that are radically different from the rest—which pull the value of the mean toward these values. We discuss more about outliers later in this chapter. EXAMPLE 4.14
Computing the Mean Cost per Order
In the Purchase Orders database, suppose that we are interested in finding the mean cost per order. We calculate the mean cost per order by summing the values in column G and then dividing by the number of observations. Using equation (4.5), note that x1 = $2,700, x2 = $19,250, and so on, and n = 94. The sum of these order costs is $2,471,760. Therefore, the mean cost per order is
▶▶ Figure
$2,471,760>94 = $26,295.32. We show these calculations in a separate worksheet, Mean, in the Purchase Orders Excel workbook. A portion of this worksheet in splitscreen mode is shown in Figure 4.21. Alternatively, we used the Excel function = AVERAGE(B2:B95) in this worksheet to arrive at the same value. We encourage you to study the calculations and formulas used.
4.21
Excel Calculations of Mean Cost per Order
Median The measure of location that specifies the middle value when the data are arranged from least to greatest is the median. Half the data are below the median, and half the data are above it. For an odd number of observations, the median is the middle of the sorted numbers. For an even number of observations, the median is the mean of the two middle numbers. We could use the Sort option in Excel to rankorder the data and then determine the median. The Excel function MEDIAN(data range) could also be used. The median is meaningful for ratio, interval, and ordinal data. As opposed to the mean, the median is not affected by outliers. EXAMPLE 4.15 Finding the Median Cost per Order In the Purchase Orders database, sort the data in column G from smallest to largest. Since we have 94 observations, the median is the average of the 47th and 48th observations. You should verify that the 47th sorted observation is $15,562.50 and the 48th observation is $15,750. Taking the average of these two values results in the median value
of ($15,562.5 + $15,750)>2 = $15,656.25. Thus, we may conclude that the costs of half the orders were less than $15,656.25 and half were above this amount. In this case, the median is not very close in value to the mean. These calculations are shown in the worksheet Median in the Purchase Orders Excel workbook, as shown in Figure 4.22.
Chapter 4 Descriptive Statistics
▶▶ Figure
163
4.22
Excel Calculations for Median Cost per Order
Mode A third measure of location is the mode. The mode is the observation that occurs most frequently. The mode is most useful for data sets that contain a relatively small number of unique values. For data sets that have few repeating values, the mode does not provide much practical value. You can easily identify the mode from a frequency distribution by identifying the value having the largest frequency or from a histogram by identifying the highest bar. You may also use the Excel function MODE.SNGL(data range). For frequency distributions and histograms of grouped data, the mode is the group with the greatest frequency. Some data sets have multiple modes; to identify these, you can use the Excel function MODE.MULT(data range), which returns an array of modal values. EXAMPL E 4.16 Finding the Mode In the Purchase Orders database, from the frequency distribution and histogram for A/P terms in Figure 4.10, we see that the greatest frequency corresponds to a value of 30 months; this is also the highest bar in the histogram.
Therefore, the mode is 30 months. For the grouped frequency distribution and histogram of the cost per order variable in Figure 4.12, we see that the mode corresponds to the group between $0 and $13,000.
Midrange A fourth measure of location that is used occasionally is the midrange. This is simply the average of the largest and smallest values in the data set. EXAMPL E 4.17 Computing the Midrange We may identify the minimum and maximum values using the Excel functions MIN and MAX or sort the data and find them easily. For the cost per order data, the minimum value
is $68.78 and the maximum value is $127,500. Thus, the midrange is ($127,500 + $68.78)>2 = $63,784.39.
Caution must be exercised when using the midrange because extreme values easily distort the result, as this example illustrates. This is because the midrange uses only two pieces of data, whereas the mean uses all the data; thus, it is usually a much rougher estimate than the mean and is often used for only small sample sizes.
Using Measures of Location in Business Decisions Because everyone is so familiar with the concept of the average in daily life, managers often use the mean inappropriately in business when other statistical information should be considered. The following hypothetical example, which was based on a real situation, illustrates this.
164
Chapter 4 Descriptive Statistics
EXAMPLE 4.18 Quoting Computer Repair Times The Excel file Computer Repair Times provides a sample of the times it took to repair and return 250 computers to customers who used the repair services of a national electronics retailer. Computers are shipped to a central facility, where they are repaired and then shipped back to the stores for customer pickup. The mean, median, and mode are all very close and show that the typical repair time is about two weeks (see Figure 4.23). So you might think that if a customer brought in a computer for repair, it would be reasonable to quote a repair time of two weeks. What would happen if the stores quoted all customers a time of two weeks? Clearly about half the customers would be upset because their computers would not be completed by this time. Figure 4.24 shows a portion of the frequency distribution and histogram for these repair times (see the
Histogram tab in the Excel file). We see that the longest repair time took almost six weeks. So should the company give customers a guaranteed repair time of six weeks? They probably wouldn’t have many customers because few would want to wait that long. Instead, the frequency distribution and histogram provide insight into making a more rational decision. You may verify that 90% of the time, repairs are completed within 21 days; on the rare occasions that it takes longer, it generally means that technicians had to order and wait for a part. So it would make sense to tell customers that they could probably expect their computers back within two to three weeks and inform them that it might take longer if a special part was needed.
From this example, we see that using frequency distributions, histograms, and percentiles can provide more useful information than simple measures of location. This leads us to introduce ways of quantifying variability in data, which we call measures of dispersion. ▶▶ Figure
4.23
Measures of Location for Computer Repair Times
▲▲ Figure
4.24
Frequency Distribution and Histogram for Computer Repair Times
Chapter 4 Descriptive Statistics
165
Measures of Dispersion: Range, Interquartile Range, Variance, and Standard Deviation Dispersion refers to the degree of variation in the data, that is, the numerical spread (or compactness) of the data. Several statistical measures characterize dispersion: the range, interquartile range, variance, and standard deviation. Range The range is the easiest to compute, and is the difference between the maximum value and the minimum value in the data set. Although Excel does not provide a function for the range, it can be computed easily by the formula =MAX(data range)  MIN(data range). Like the midrange, the range is affected by outliers and, thus, is often only used for very small data sets. EXAMPL E 4.19 Computing the Range For the cost per order data in the Purchase Orders database, the minimum value is $68.78 and the
maximum value is $127,500. Thus, the range is $127,500  $68.78 = $127,431.22.
Interquartile Range The difference between the first and third quartiles, Q3  Q1, is often called the interquartile range (IQR), or the midspread. This includes only the middle 50% of the data and, therefore, is not influenced by extreme values. Thus, it is sometimes used as an alternative measure of dispersion. EXAMPL E 4.20 Computing the Interquartile Range For the cost per order data in the Purchase Orders database, we identified the first and third quartiles as Q1 = $6,757.81 and Q3 = $27,593.75 in Example 4.12. Thus, IQR = $27,593.75  $6,757.81 = $20,835.94. Therefore, the middle 50% of the data are concentrated
over a relatively small range of $20,835.94. Note that the upper 25% of the data span the range from $27,593.75 to $127,500, indicating that high costs per order are spread out over a large range of $99,906.25.
Variance A more commonly used measure of dispersion is the variance, whose computation depends on all the data. The larger the variance, the more the data are spread out from the mean and the more variability one can expect in the observations. The formula used for calculating the variance is different for populations and samples. The formula for the variance of a population is 2 a 1xi  m2 N
s2 =
i=1
(4.7) N where xi is the value of the ith item, N is the number of items in the population, and m is the population mean. Essentially, the variance is the average of the squared deviations of the observations from the mean. A significant difference exists between the formulas for computing the variance of a population and that of a sample. The variance of a sample is calculated using the formula
2 a 1xi  x2 n
s2 =
i=1
n  1
(4.8)
166
Chapter 4 Descriptive Statistics
where n is the number of items in the sample and x is the sample mean. It may seem peculiar to use a different denominator to “average” the squared deviations from the mean for populations and samples, but statisticians have shown that the formula for the sample variance provides a more accurate representation of the true population variance. We discuss this more formally in Chapter 6. For now, simply understand that the proper calculations of the population and sample variance use different denominators based on the number of observations in the data. The Excel function VAR.S(data range) may be used to compute the sample variance, s2, whereas the Excel function VAR.P(data range) is used to compute the variance of a population, s2. EXAMPLE 4.21
Computing the Variance
Figure 4.25 shows a portion of the Excel worksheet Variance in the Purchase Orders workbook. To find the variance of the cost per order using equation (4.8), we first need to calculate the mean, as was done in Example 4.14. Then for each observation, calculate the difference between the ▶▶ Figure
observation and the mean, as shown in column C. Next, square these differences, as shown in column D. Finally, add these square deviations (cell D96) and divide by n  1 = 93. This results in the variance 890,594,573.82. Alternatively, the Excel function = VAR.S(B2:B95) yields the same result.
4.25
Excel Calculations for Variance of Cost per Order
Note that the dimension of the variance is the square of the dimension of the observations. So for example, the variance of the cost per order is not expressed in dollars, but rather in dollars squared. This makes it difficult to use the variance in practical applications. However, a measure closely related to the variance that can be used in practical applications is the standard deviation. Standard Deviation The standard deviation is the square root of the variance. For a population, the standard deviation is computed as 2 a 1xi  m2 N
s =
and for samples, it is
R
i=1
N
2 a 1xi  x2
( 4.9)
n
s =
i=1
R
n  1
( 4.10)
Chapter 4 Descriptive Statistics
167
The Excel function STDEV.P(data range) calculates the standard deviation for a population 1s2; the function STDEV.S(data range) calculates it for a sample (s).
EXAMPL E 4.22 Computing the Standard Deviation We may use the same worksheet calculations as in Example 4.21. All we need to do is to take the square root of the computed variance to find the standard deviation. Thus, the standard deviation of the cost per order is
2890,594,573.82 = $29,842.8312. Alternatively, we could use the Excel function = STDEV.S(B2:B95) to find the same value.
The standard deviation is generally easier to interpret than the variance because its units of measure are the same as the units of the data. Thus, it can be more easily related to the mean or other statistics measured in the same units. The standard deviation is a popular measure of risk, particularly in financial analysis, because many people associate risk with volatility in stock prices. The standard deviation measures the tendency of a fund’s monthly returns to vary from their longterm average (as Fortune stated in one of its issues, “. . . standard deviation tells you what to expect in the way of dips and rolls. It tells you how scared you’ll be.”).4 For example, a mutual fund’s return might have averaged 11% with a standard deviation of 10%. Thus, about twothirds of the time, the annualized monthly return was between 1% and 21%. By contrast, another fund’s average return might be 14% but have a standard deviation of 20%. Its returns would have fallen in a range of 6% to 34% and, therefore, it is more risky. Many financial Web sites, such as IFA.com and Morningstar.com, provide standard deviations for market indexes and mutual funds. To illustrate risk, the Excel file Closing Stock Prices (see Figure 4.26) lists daily closing prices for four stocks and the Dow Jones Industrial Average index over a onemonth period. The average closing prices for Intel (INTC) and General Electric (GE) are quite ▶▶ Figure
4.26
Excel File Closing Stock Prices
4
Fortune, 1999 Investor’s Guide (December 21, 1998).
168
Chapter 4 Descriptive Statistics
similar, $18.81 and $16.19, respectively. However, the standard deviation of Intel’s price over this time frame was $0.50, whereas GE’s was $0.35. GE had less variability and, therefore, less risk. A larger standard deviation implies that while a greater potential of a higher return exists, there is also greater risk of realizing a lower return. Many investment publications and Web sites provide standard deviations of stocks and mutual funds to help investors assess risk in this fashion. We learn more about risk in other chapters.
Chebyshev’s Theorem and the Empirical Rules One of the more important results in statistics is Chebyshev’s theorem, which states that for any set of data, the proportion of values that lie within k standard deviations 1k 7 12 of the mean is at least 1  1>k 2. Thus, for k = 2, at least 3/4, or 75%, of the data lie within two standard deviations of the mean; for k = 3, at least 8/9, or 89%, of the data lie within three standard deviations of the mean. We can use these values to provide a basic understanding of the variation in a set of data using only the computed mean and standard deviation. EXAMPLE 4.23 Applying Chebyshev’s Theorem For the cost per order data in the Purchase Orders database, a twostandarddeviation interval around the mean is [$33,390.34, $85,980.98]. If we count the number of observations within this interval, we find that 89 of 94, or 94.7%, fall within two standard deviations
of the mean. A threestandarddeviation interval is [$63,233.17, $115,823.81], and we see that 92 of 94, or 97.9%, fall in this interval. Both percentages are above the thresholds of 75% and 89% in Chebyshev’s theorem.
For many data sets encountered in practice, such as the cost per order data, the percentages are generally much higher than what Chebyshev’s theorem specifies. These are reflected in what are called the empirical rules: 1. Approximately 68% of the observations will fall within one standard deviation of the mean, or between x  s and x + s. 2. Approximately 95% of the observations will fall within two standard deviations of the mean, or within x { 2s. 3. Approximately 99.7% of the observations will fall within three standard deviations of the mean, or within x { 3s. We see that the cost per order data reflect these empirical rules rather closely. Depending on the data and the shape of the frequency distribution, the actual percentages may be higher or lower. To illustrate the empirical rules, suppose that a retailer knows that on average, an order is delivered by standard ground transportation in eight days with a standard deviation of one day. With the first empirical rule, about 68% of the observations will fall between seven and nine days. Using the second empirical rule, 95% of the observations will fall between six and ten days. Finally, using the third rule, almost 100% of the observations will fall between 5 and 11 days. These rules can be used to predict future observations, assuming that the sample data are representative of the future. Generally, rules 2 and 3 are commonly used. Note, however, that using rule 2, there is a 5% chance that an observation will fall outside the predicted range. Since the third empirical rule generally covers nearly all of the observations, it provides a better prediction and reduces the risk that a future observation will fall outside of the predicted range. As another example, it is important to ensure that the output from a manufacturing process meets the specifications that engineers and designers require. The dimensions for
Chapter 4 Descriptive Statistics
169
a typical manufactured part are usually specified by a target, or ideal, value as well as a tolerance, or “fudge factor,” that recognizes that variation will exist in most manufacturing processes due to factors such as materials, machines, work methods, human performance, and environmental conditions. For example, a part dimension might be specified as 5.00 { 0.2 cm. This simply means that a part having a dimension between 4.80 and 5.20 cm will be acceptable; anything outside of this range would be classified as defective. To measure how well a manufacturing process can achieve the specifications, we usually take a sample of output, measure the dimension, compute the total variation using the third empirical rule (that is, estimate the total variation by six standard deviations), and then compare the result to the specifications by dividing the specification range by the total variation. The result is called the process capability index, denoted as Cp:
Cp =
Upper Specification  Lower Specification (4.11) Total Variation
Manufacturers use this index to evaluate the quality of their products and determine when they need to make improvements in their processes. EXAMPL E 4.24 Using Empirical Rules to Measure the Capability of a
Manufacturing Process Figure 4.27 shows a portion of the data collected from a manufacturing process for a part whose dimensions are specified as 5.00 { 0.2 centimeters. These are provided in the Excel workbook Manufacturing Measurements. The mean and standard deviation are first computed in cells J3 and J4 using the Excel AVERAGE and STDEV.S functions (these functions work correctly whether the data are arranged in a single column or in a matrix form). Using the third empirical rule, the total variation is then calculated as the mean plus or minus three standard deviations. In cell J14, Cp is calculated using equation (4.11). A Cp value less than 1.0 is not good; it means that the variation in the process is wider than the specification limits, signifying that some of the parts will not meet the specifications. In practice, many manufacturers want to have Cp values of at least 1.5.
▲▲ Figure
4.27
Calculation of Cp Index
Figure 4.28 shows a frequency distribution and histogram of these data (worksheet Histogram in the Manufacturing Measurements workbook). Note that the bin values represent the upper limits of the groupings in the histogram; thus, three observations fell at or below 4.8, the lower specification limit. In addition, five observations exceeded the upper specification limit of 5.2. Therefore, 8 of the 200 observations, or 4%, were actually defective, and 96% were acceptable. Although this doesn’t meet the empirical rule exactly, you must remember that we are dealing with sample data. Other samples from the same process would have different characteristics, but overall, the empirical rule provides a good estimate of the total variation in the data that we can expect from any sample.
170
Chapter 4 Descriptive Statistics
▶▶ Figure
4.28
Frequency Distribution and Histogram of Manufacturing Measurements
Standardized Values (ZScores) A standardized value, commonly called a zscore, provides a relative measure of the distance an observation is from the mean, which is independent of the units of measurement. The zscore for the ith observation in a data set is calculated as follows:
zi =
xi  x ( 4.12) s
We subtract the sample mean from the ith observation, xi, and divide the result by the sample standard deviation. In formula (4.12), the numerator represents the distance that xi is from the sample mean; a negative value indicates that xi lies to the left of the mean, and a positive value indicates that it lies to the right of the mean. By dividing by the standard deviation, s, we scale the distance from the mean to express it in units of standard deviations. Thus, a zscore of 1.0 means that the observation is one standard deviation to the right of the mean; a zscore of 1.5 means that the observation is 1.5 standard deviations to the left of the mean. Thus, even though two data sets may have different means and standard deviations, the same zscore means that the observations have the same relative distance from their respective means. Zscores can be computed easily on a spreadsheet; however, Excel has a function that calculates it directly, STANDARDIZE(x, mean, standard_dev).
EXAMPLE 4.25
Computing ZScores
Figure 4.29 shows the calculations of zscores for a portion of the cost per order data. This worksheet may be found in the Purchase Orders workbook as zscores. In cells B97 and B98, we compute the mean and standard deviation using the Excel AVERAGE and STDEV.S functions. In column C, we could use either formula (4.12) or the Excel STANDARDIZE function. For example, the formula in cell C2 is = (B2 $B$97)>$B$98, but it could also be calculated
as = STANDARDIZE(B2,$B$97,$B$98). Thus, the first observation, $82,875, is 1.90 standard deviations above the mean, whereas observation 94 ($68.75) is 0.88 standard deviations below the mean. Only two observations are more than three standard deviations above the mean. We saw this in Example 4.23, when we applied Chebyshev’s theorem to the data.
Chapter 4 Descriptive Statistics
▶▶ Figure
171
4.29
Computing ZScores for Cost per Order Data
Coefficient of Variation The coefficient of variation (CV) provides a relative measure of the dispersion in data relative to the mean and is defined as
CV =
Standard Deviation (4.13) Mean
Sometimes the coefficient of variation is multiplied by 100 to express it as a percent. This statistic is useful when comparing the variability of two or more data sets when their scales differ. The coefficient of variation provides a relative measure of risk to return. The smaller the coefficient of variation, the smaller the relative risk is for the return provided. The reciprocal of the coefficient of variation, called return to risk, is often used because it is easier to interpret. That is, if the objective is to maximize return, a higher returntorisk ratio is often considered better. A related measure in finance is the Sharpe ratio, which is the ratio of a fund’s excess returns (annualized total returns minus Treasury bill returns) to its standard deviation. If several investment opportunities have the same mean but different variances, a rational (riskaverse) investor will select the one that has the smallest variance.5 This approach to formalizing risk is the basis for modern portfolio theory, which seeks to construct minimumvariance portfolios. As Fortune magazine once observed, “It’s not that risk is always bad. . . . It’s just that when you take chances with your money, you want to be paid for it.”6 One practical application of the coefficient of variation is in comparing stock prices.
EXAMPL E 4.26 Applying the Coefficient of Variation If examining only the standard deviations in the Closing Stock Prices worksheet, we might conclude that IBM is more risky than the other stocks. However, the mean stock price of IBM is much greater than the other stocks. Thus, comparing standard deviations directly provides little information. The coefficient of variation provides a more comparable measure. Figure 4.30 shows the calculations of
5 David 6
the coefficients of variation for these variables. For IBM, the CV is 0.025; for Intel, 0.027; for Cisco, 0.024; for GE, 0.022; and for the DJIA, 0.016. We see that the coefficients of variation of the stocks are not very different; in fact, Intel is just slightly more risky than IBM relative to its average price. However, an index fund based on the Dow would be less risky than any of the individual stocks.
G. Luenberger, Investment Science (New York: Oxford University Press, 1998). Fortune, 1999 Investor’s Guide (December 21, 1998).
172
Chapter 4 Descriptive Statistics
▶▶ Figure
4.30
Calculating Coefficients of Variation for Closing Stock Prices
Measures of Shape Histograms of sample data can take on a variety of different shapes. Figure 4.31 shows the histograms for cost per order and A/P terms that we created for the Purchase Orders data. The histogram for A/P terms is relatively symmetric, having its modal value in the middle and falling away from the center in roughly the same fashion on either side. However, the cost per order histogram is asymmetrical, or skewed; that is, more of the mass is concentrated on one side, and the distribution of values “tails off” to the other. Those that have more mass on the left and tail off to the right, like this example, are called positively skewed; those that have more mass on the right and tail off to the left are said to be negatively skewed. Skewness describes the lack of symmetry of data. The coefficient of skewness (CS) measures the degree of asymmetry of observations around the mean. The coefficient of skewness for a population is computed as
CS =
1 N 1xi  m2 3 N ia =1 s3
( 4.14)
It can be computed using the Excel function SKEW.P(data range). For sample data, the formula is different, but you may calculate it using the Excel function SKEW(data range). If CS is positive, the distribution of values is positively skewed; if negative, it is negatively skewed. The closer CS is to zero, the less the degree of skewness. A coefficient of skewness greater than 1 or less than 1 suggests a high degree of skewness. A value between 0.5 and 1 or between 0.5 and 1 represents moderate skewness. Coefficients between 0.5 and 0.5 indicate relative symmetry. EXAMPLE 4.27
Measuring Skewness
Using the Excel function SKEW in the Purchase Orders database, the coefficients of skewness for the cost per order and A/P terms data are calculated as CS (cost per order) = 1.66 CS (A>P terms) = 0.60
This tells us that the cost per order data are highly positively skewed, whereas the A/P terms data are more symmetric. These are evident from the histograms in Figure 4.31.
Histograms that have only one “peak” are called unimodal. (If a histogram has exactly two peaks, we call it bimodal. This often signifies a mixture of samples from different populations.) For unimodal histograms that are relatively symmetric, the mode is a fairly good estimate of the mean. For example, the mode for the A/P terms data is clearly 30 months; the mean is 30.638 months. On the other hand, for the cost per order data,
▲▲ Figure
Chapter 4 Descriptive Statistics
173
4.31
Histograms of Cost per Order and A/P Terms ▶▶ Figure
4.32
Characteristics of Skewed Distributions
Mean
Median
Mode
Mode
Median
Mean
the mode occurs in the group (0, 13,000). The midpoint of the group, $6,500, which can be used as a numerical estimate of the mode, is not very close at all to the true mean of $26,295.32. The high level of skewness pulls the mean away from the mode. Comparing measures of location can sometimes reveal information about the shape of the distribution of observations. For example, if the distribution was perfectly symmetrical and unimodal, the mean, median, and mode would all be the same. If it was negatively skewed, we would generally find that mean 6 median 6 mode, whereas a positive skewness would suggest that mode 6 median 6 mean (see Figure 4.32). Kurtosis refers to the peakedness (that is, high, narrow) or flatness (that is, short, flattopped) of a histogram. The coefficient of kurtosis (CK) measures the degree of kurtosis of a population and is computed as 1 N 1xi  m2 4 N ia =1
(4.15) s4 Distributions with values of CK less than 3 are more flat, with a wide degree of dispersion; those with values of CK greater than 3 are more peaked, with less dispersion. Excel computes kurtosis differently; the function KURT (data range) computes “excess kurtosis” for sample data (Excel does not have a corresponding function for a population), which is CK − 3. Thus, to interpret kurtosis values in Excel, distributions with values less than 0 are more flat, while those with values greater than 0 are more peaked. Skewness and kurtosis can help provide more information to evaluate risk than just using the standard deviation. For example, both a negatively and positively skewed distribution may have the same standard deviation, but clearly if the objective is to achieve a high return, the negatively skewed distribution will have higher probabilities of larger returns. The higher the kurtosis, the more area the histogram has in the tails rather than in the middle. This can indicate a greater potential for extreme and possibly catastrophic outcomes.
CK =
174
Chapter 4 Descriptive Statistics
Excel Descriptive Statistics Tool Excel provides a useful tool for basic data analysis, Descriptive Statistics, which provides a summary of numerical statistical measures that describe location, dispersion, and shape for sample data (not a population). Click on Data Analysis in the Analysis group under the Data tab in the Excel menu bar. Select Descriptive Statistics from the list of tools. The Descriptive Statistics dialog shown in Figure 4.33 will appear. You need to enter only the range of the data, which must be in a single row or column. If the data are in multiple columns, the tool treats each row or column as a separate data set, depending on which you specify. This means that if you have a single data set arranged in a matrix format, you would have to stack the data in a single column before applying the Descriptive Statistics tool. Check the box Labels in First Row if labels are included in the input range. You may choose to save the results in the current worksheet or in a new one. For basic summary statistics, check the box Summary statistics; you need not check any others. EXAMPLE 4.28
Using the Descriptive Statistics Tool
We will apply the Descriptive Statistics tool to the cost per order and A/P terms data in columns G and H of the Purchase Orders database. The results are provided in the Descriptive Statistics worksheet in the Purchase Orders
▶▶ Figure
4.33
Descriptive Statistics Dialog
▶▶ Figure
4.34
Purchase Orders Data Descriptive Statistics Summary
workbook and are shown in Figure 4.34. The tool provides all the measures we have discussed as well as the standard error, which we discuss in Chapter 6, along with the minimum, maximum, sum, and count.
Chapter 4 Descriptive Statistics
175
One important point to note about the use of the tools in the Analysis Toolpak versus Excel functions is that while Excel functions dynamically change as the data in the spreadsheet are changed, the results of the Analysis Toolpak tools do not. For example, if you compute the average value of a range of numbers directly using the function AVERAGE(range), then changing the data in the range will automatically update the result. However, you would have to rerun the Descriptive Statistics tool after changing the data. CHECK YOUR UNDERSTANDING 1. Explain the difference between a population and a sample. 2. List and explain the common measures of location used in statistics. 3. List and explain the common measures of dispersion used in statistics. 4. How are Chebyshev’s theorem and the empirical rules useful in business? 5. What is a standardized value (zscore)? 6. Explain the value of the coefficient of variation. 7. Define the common measures of shape used in statistics. 8. What information does the Excel Descriptive Statistics tool provide?
Computing Descriptive Statistics for Frequency Distributions When data are summarized in a frequency distribution, we can use the frequencies to compute the mean and variance. The following formulas provide a way to calculate the mean and variance of data expressed as a frequency distribution. The mean of a population expressed as a frequency distribution may be computed using the formula a fi xi N
m =
i=1
N
(4.16)
For samples, the formula is similar: a fi xi n
x =
i=1
n
(4.17)
where fi is the frequency of observation i. Essentially, we multiply the frequency by the value of observation i, add them up, and divide by the number of observations. We may use similar formulas to compute the population variance 2 a fi 1xi  m2 N
s2 =
i=1
N
(4.18)
176
Chapter 4 Descriptive Statistics
and sample variance 2 a fi 1xi  x2 n
s2 =
i=1
n  1
( 4.19)
To find the standard deviation, take the square root of the variance, as we did earlier. Note the similarities between these formulas and formulas (4.16) and (4.17). In multiplying the values by the frequency, we are essentially adding the same values fi times. So they really are the same formulas, just expressed differently. EXAMPLE 4.29 Computing Statistical Measures from Frequency Distributions In Example 4.4, we constructed a frequency distribution for A/P terms in the Purchase Orders database. Figure 4.35 shows the calculations of the mean and sample variance using formulas (4.17) and (4.19) for the frequency distribution. In column C, we multiply the frequency by the value of
the observations [the numerator in formula (4.17)] and then divide by n, the sum of the frequencies in column B, to find the mean. Columns D, E, and F provide the calculations needed to find the variance. We divide the sum of the data in column F by n  1 = 93 to find the variance.
In some situations, data may already be grouped in a frequency distribution, and we may not have access to the raw data. This is often the case when extracting information from government databases such as the Census Bureau or Bureau of Labor Statistics. In these situations, we cannot compute the mean or variance using the standard formulas. If the data are grouped into k cells in a frequency distribution, we can use modified versions of these formulas to estimate the mean and variance by replacing xi with a representative value (such as the midpoint, M) for all the observations in each cell group and summing over all groups. The corresponding formulas follow. In these formulas, k is the number of groups and Mi is the midpoint of group i. Population mean estimate for grouped data: a fiMi k
m =
i=1
N
( 4.20)
Sample mean estimate for grouped data: a fi Mi k
▶▶ Figure
4.35
Calculations of Mean and Variance Using a Frequency Distribution
x =
i=1
n
( 4.21)
Chapter 4 Descriptive Statistics
▶▶ FIGURE
177
4.36
Grouped Frequency Calculations for Cost per Order
Population variance estimate for grouped data: 2 a fi 1Mi  m2 k
s2 =
i=1
N
(4.22)
Sample variance estimate for grouped data: 2 a fi 1Mi  x2 k
s2 =
i=1
n  1
(4.23)
EXAMPL E 4.30 Computing Descriptive Statistics for
a Grouped Frequency Distribution In Figure 4.11, we illustrated a grouped frequency distribution for cost per order in the Purchase Orders database. Figure 4.36 shows the calculations of the mean and sample variance using formulas (4.21) and (4.23). It is
important to understand that because we have not used all the original data in computing these statistics, they are only estimates of the true values.
CHECK YOUR UNDERSTANDING 1. Explain the process of computing the mean and variance from a frequency distribution and from a grouped frequency distribution. 2. Use Excel to find the mean and variance for the grouped frequency distribution in Figure 4.12.
Descriptive Statistics for Categorical Data: The Proportion Statistics such as means and variances are not appropriate for categorical data. Instead, we are generally interested in the fraction of data that have a certain characteristic. The formal statistical measure is called the proportion, usually denoted by p. The proportion is computed using the formula
p = x>n (4.24)
where x is the number of observations having a certain characteristic and n is the sample size. Note that proportions are analogous to relative frequencies for categorical data.
178
Chapter 4 Descriptive Statistics
Proportions are key descriptive statistics for categorical data, such as demographic data from surveys, defects or errors in qualitycontrol applications, or consumer preferences in market research. EXAMPLE 4.31
Computing a Proportion
In the Purchase Orders database, column A lists the name of the supplier for each order. We may use the Excel function = COUNTIF(data range, criteria) to count the number of observations meeting specified characteristics. For instance, to find the number of orders
placed with Spacetime Technologies, we used the function = COUNTIF(A4:A97, “Spacetime Technologies”). This returns a value of 12. Because 94 orders were placed, the proportion of orders placed with Spacetime Technologies is p = 12>94 = 0.128.
It is important to realize that proportions are numbers between 0 and 1. Although we often convert these to percentages—for example, 12.8% of orders were placed with Spacetime Technologies in the last example—we must be careful to use the decimal expression of a proportion when statistical formulas require it. CHECK YOUR UNDERSTANDING 1. What is a proportion, and how is it computed? 2. Find the proportion of orders placed for boltnut packages in the Purchase Orders database.
Statistics in PivotTables We introduced PivotTables in Chapter 2. PivotTables also have the functionality to calculate many basic statistical measures from the data summaries. If you look at the Value Field Settings dialog shown in Figure 4.37, you can see that you can calculate the average, standard deviation, and variance of a value field. ▶▶ Figure
4.37
Value Field Settings Dialog
Chapter 4 Descriptive Statistics
▶▶ Figure
179
4.38
PivotTable for Average Checking and Savings Account Balances by Job
EXAMPL E 4.32
Statistical Measures in PivotTables
In the Credit Risk Data Excel file, suppose that we want to find the average amount of money in checking and savings accounts by job classification. Create a PivotTable, and in the PivotTable Field List, move Job to the Row Labels field and Checking and Savings to the Values field. Then change the field settings from “Sum of Checking”
and “Sum of Savings” to the averages. The result is shown in Figure 4.38; we have also formatted the values as currency using the Number Format button in the dialog. In a similar fashion, you could find the standard deviation or variance of each group by selecting the appropriate field settings.
CHECK YOUR UNDERSTANDING 1. What statistical information can be displayed in a PivotTable? 2. Use a PivotTable to find the standard deviation of the amount of money in checking and savings accounts for each job classification in the Credit Risk Data Excel file.
Measures of Association Two variables have a strong statistical relationship with one another if they appear to move together. We see many examples on a daily basis; for instance, attendance at baseball games is often closely related to the win percentage of the team, and ice cream sales likely have a strong relationship with daily temperature. We can examine relationships between two variables visually using scatter charts, which we introduced in Chapter 3. When two variables appear to be related, you might suspect a causeandeffect relationship. Sometimes, however, statistical relationships exist even though a change in one variable is not caused by a change in the other. For example, the New York Times reported a strong statistical relationship between the golf handicaps of corporate CEOs and their companies’ stock market performance over three years. CEOs who were betterthanaverage golfers were likely to deliver aboveaverage returns to shareholders.7 Clearly, the ability to golf would not cause better business performance. Therefore, you must be cautious in drawing inferences about causal relationships based solely on statistical relationships. (On the other hand, you might want to spend more time out on the practice range!) Understanding the relationships between variables is extremely important in making good business decisions, particularly when causeandeffect relationships can be justified. When a company understands how internal factors such as product quality, employee training, and pricing affect such external measures as profitability and customer satisfaction, it can make better decisions. Thus, it is helpful to have statistical tools for measuring these relationships. 7
Adam Bryant, “CEOs’ Golf Games Linked to Companies’ Performance,” Cincinnati Enquirer, June 7, 1998, El.
180
Chapter 4 Descriptive Statistics
▲▲ Figure
4.39
Portion of Excel File Colleges and Universities
The Excel file Colleges and Universities, a portion of which is shown in Figure 4.39, contains data from 49 top liberal arts and research universities across the United States. Several questions might be raised about statistical relationships among these variables. For instance, does a higher percentage of students in the top 10% of their high school class suggest a higher graduation rate? Is acceptance rate related to the amount spent per student? Do schools with lower acceptance rates tend to accept students with higher SAT scores? Questions such as these can be addressed by computing statistical measures of association between the variables.
Covariance Covariance is a measure of the linear association between two variables, X and Y. Like the variance, different formulas are used for populations and samples. Computationally, covariance of a population is the average of the products of deviations of each observation from its respective mean: a 1xi  mx 21yi  my 2 N
cov(X, Y) =
i=1
N
( 4.25)
To better understand the covariance, let us examine formula (4.25). The covariance between X and Y is the average of the product of the deviations of each pair of observations from their respective means. Suppose that large (small) values of X are generally associated with large (small) values of Y. Then, in most cases, both xi and yi are either above or below their respective means. If so, the product of the deviations from the means will be a positive number and when added together and averaged will give a positive value for the covariance. On the other hand, if small (large) values of X are associated with large (small) values of Y, then one of the deviations from the mean will generally be negative while the other is positive. When multiplied together, a negative value results, and the value of the covariance will be negative. Thus, the larger the absolute value of the covariance, the higher the degree of linear association between the two variables. The sign of the covariance tells us whether there is a direct relationship (that is, one variable increases as the other increases) or an inverse relationship (that is, one variable increases while the other decreases, or vice versa). We can generally identify the strength of any linear association between two variables and the sign of the covariance by constructing a scatter diagram. The Excel function COVARIANCE.P(array1, array2) computes the covariance of a population.
Chapter 4 Descriptive Statistics
181
The sample covariance is computed as a 1xi  x21yi  y2 n
cov(X, Y) =
i=1
n  1
(4.26)
Similar to the sample variance, note the use of n  1 in the denominator. The Excel function COVARIANCE.S(array1, array2) computes the covariance of a sample. EXAMPL E 4.33
Computing the Covariance
Figure 4.40 shows a scatter chart of graduation rates (Y variable) versus median SAT scores (X variable) for the Colleges and Universities data. It appears that as the median SAT scores increase, the graduation rate also increases; thus, we would expect to see a positive covariance. Figure
4.41 shows the calculations using formula (4.26); these are provided in the worksheet Covariance Calculations in the Colleges and Universities Excel workbook. The Excel function = COVARIANCE.S(B2:B50,C2:C50) in cell F55 verifies the calculations.
Correlation The numerical value of the covariance is generally difficult to interpret because it depends on the units of measurement of the variables. For example, if we expressed the graduation rate as a true proportion rather than as a percentage in the previous example, the numerical value of the covariance would be smaller, although the linear association between the variables would be the same. Correlation is a measure of the linear relationship between two variables, X and Y, which does not depend on the units of measurement. Correlation is measured by the correlation coefficient, also known as the Pearson product moment correlation coefficient. The correlation coefficient for a population is computed as
▶▶ Figure
4.40
Scatter Chart of Graduation Rate Versus Median SAT
rXY =
cov(X, Y) (4.27) sX sY
182
Chapter 4 Descriptive Statistics
▶▶ Figure
4.41
Covariance Calculations for Graduation Rate and Median SAT
where sX is the standard deviation of X, and sY is the standard deviation of Y. By dividing the covariance by the product of the standard deviations, we are essentially scaling the numerical value of the covariance to a number between 1 and 1. In a similar fashion, the sample correlation coefficient is computed as
rXY =
cov(X, Y) ( 4.28) sXsY
Excel’s CORREL(array1, array2) function computes the correlation coefficient of two data arrays. A correlation of 0 indicates that the two variables have no linear relationship to each other. Thus, if one changes, we cannot reasonably predict what the other variable might do. A positive correlation coefficient indicates a linear relationship when one variable increases while the other also increases. A negative correlation coefficient indicates a linear relationship when one variable that increases while the other decreases. In economics, for instance, a priceelastic product has a negative correlation between price and sales; as price increases, sales decrease, and vice versa. Visualizations of correlation are illustrated in Figure 4.42. Note that although Figure 4.42(d) shows a clear relationship between the variables, the relationship is not linear and the correlation is zero. ▶▶ Figure
4.42
Examples of Correlation Y 15
10
5
0
5
10
15
Y
X
X (a) Positive Correlation
Y
(b) Negative Correlation
Y
X (c) No Correlation
X (d) A Nonlinear Relationship with No Linear Correlation
Chapter 4 Descriptive Statistics
▲▲ Figure
183
4.43
Correlation Calculations for Graduation Rate and Median SAT
EXAMPL E 4.34
Computing the Correlation Coefficient
Figure 4.43 shows the calculations for computing the sample correlation coefficient for the graduation rate and median SAT variables in the Colleges and Universities data file. We first compute the standard deviation of each
variable in cells B52 and C52 and then divide the covariance by the product of these standard deviations in cell F54. Cell F56 shows the same result using the Excel function = CORREL(B2:B50,C2:C50).
When using the CORREL function, it does not matter if the data represent samples or populations. In other words, CORREL(array1, array2) =
COVARIANCE.P1array1, array22 STDEV.P1array12 * STDEV.P1array22
CORREL(array1, array2) =
COVARIANCE.S1array1, array22 STDEV.S1array12 * STDEV.S1array22
and
For instance, in Example 4.34, if we assume that the data are populations, we find that the population standard deviation for X is 7.372 and the population standard deviation for Y is 62.034 (using the function STDEV.P). By dividing the population covariance, 257.995 (using the function COVARIANCE.P), by the product of these standard deviations, we find that the correlation coefficient is still 0.564 as computed by the CORREL function.
Excel Correlation Tool The Data Analysis Correlation tool computes correlation coefficients for more than two arrays. Select Correlation from the Data Analysis tool list. The dialog is shown in Figure 4.44. You need to input only the range of the data (which must be in contiguous columns; if not, you must move them in your worksheet), specify whether the data are grouped by rows or columns (most applications will be grouped by columns), and indicate whether the first row contains data labels. The output of this tool is a matrix giving the correlation between each pair of variables. This tool provides the same output as the CORREL function for each pair of variables.
184
Chapter 4 Descriptive Statistics
▶▶ Figure
4.44
Excel Correlation Tool Dialog
▶▶ Figure
4.45
Correlation Results for Colleges and Universities Data
EXAMPLE 4.35
Using the Correlation Tool
The correlation matrix among all the variables in the Colleges and Universities data file is shown in Figure 4.45. None of the correlations are very strong. The moderate positive correlation between the graduation rate and SAT scores indicates that schools with higher median SATs have higher graduation rates. We see a moderate negative correlation between acceptance rate and graduation rate,
indicating that schools with lower acceptance rates have higher graduation rates. We also see that the acceptance rate is also negatively correlated with the median SAT and top 10% HS, suggesting that schools with lower acceptance rates have higher student profiles. The correlations with expenditures/student also suggest that schools with higher student profiles spend more money per student.
CHECK YOUR UNDERSTANDING 1. Explain the difference between covariance and correlation. 2. What Excel functions and tools can you use to find the covariance and correlation of populations and samples? 3. Explain how to interpret the correlation coefficient.
Outliers Earlier we noted that the mean and range are sensitive to outliers—unusually large or small values in the data. Outliers can make a significant difference in the results we obtain from statistical analyses. An important statistical question is how to identify them. The first thing to do from a practical perspective is to check the data for possible errors, such as a misplaced decimal point or an incorrect transcription to a computer file. Histograms can help to identify possible outliers visually. We might use the empirical rule and zscores to identify an outlier as one that is more than three standard deviations from the mean. We can also identify outliers based on the interquartile range. “Mild” outliers are often defined as being between 1.5 × IQR and 3 × IQR to the left of Q1 or to the right of Q3, and “extreme” outliers as more than 3 × IQR away from these quartiles. Basically, there is no standard definition of what constitutes an outlier other than an unusual observation as compared with the rest. However, it is important to try to identify outliers and determine their significance when conducting business analytic studies.
Chapter 4 Descriptive Statistics
EXAMPL E 4.36
Investigating Outliers
The Excel data file Home Market Value provides a sample of data for homes in a neighborhood (Figure 4.46). Figure 4.47 shows zscore calculations for the square feet and market value variables. None of the zscores for either of these variables exceed 3 (these calculations can be found in the worksheet Outliers in the Excel Home Market Value workbook). However, while individual variables might not exhibit outliers, combinations of them might. We see this in the scatter diagram in Figure 4.48. The last observation has a high market value ($120,700) but a relatively small house size
▶▶ Figure
4.46
Portion of Home Market Value Excel File
▶▶ Figure
4.47
Computing ZScores for Examining Outliers
▶▶ Figure
4.48
Scatter Diagram of House Size Versus Market Value
185
(1,581 square feet). The point on the scatter diagram does not seem to coincide with the rest of the data. The question is what to do with possible outliers. They should not be blindly eliminated unless there is a legitimate reason for doing so—for instance, if the last home in the Home Market Value example has an outdoor pool that makes it significantly different from the rest of the neighborhood. Statisticians often suggest that analyses should be run with and without the outliers so that the results can be compared and examined critically.
186
Chapter 4 Descriptive Statistics
CHECK YOUR UNDERSTANDING 1. What rules are typically used to identify outliers? 2. How should you handle outliers in your analysis?
Using Descriptive Statistics to Analyze Survey Data Many business applications deal with survey data. Descriptive statistics tools are extremely valuable for summarizing and analyzing survey data. Consider, for example, the data in the Excel file Insurance Survey, which is shown in Figure 4.49. We may compute various statistical measures and apply statistical tools to understand these data, such as the following: ■■ ■■ ■■
■■ ■■ ■■
▶▶ Figure
4.49
Insurance Survey Data
Frequency distributions and histograms for the ratio variables (age, years employed, and satisfaction) Descriptive statistical measures for the ratio variables using the Descriptive Statistics tool Proportions for various attributes of the categorical variables in the sample, such as the proportion of males and females, different educational levels, and marital status PivotTables that break down the averages of ratio variables by gender, education, and marital status Crosstabulations by gender and education, gender and marital status, and education and marital status Zscores for examination of potential outliers
Chapter 4 Descriptive Statistics
187
Problem 69 at the end of this chapter asks you to use these approaches to analyze these data. These tools can be supplemented by various data visualization tools that we discussed in the previous chapter. Such analysis can provide the basis for formal reports and presentations that one might use to communicate and explain results to managers. CHECK YOUR UNDERSTANDING 1. Summarize statistical tools that are typically used to analyze survey data. 2. Explain the importance of using statistics for communicating survey data.
Statistical Thinking in Business Decisions The importance of applying statistical concepts to make good business decisions and improve performance cannot be overemphasized. Statistical thinking is a philosophy of learning and action for improvement that is based on the principles that ■■ ■■ ■■
all work occurs in a system of interconnected processes, variation exists in all processes, and better performance results from understanding and reducing variation.8
Work gets done in any organization through processes—systematic ways of doing things that achieve desired results. Understanding business processes provides the context for determining the effects of variation and the proper type of action to be taken. Any process contains many sources of variation. In manufacturing, for example, different batches of material vary in strength, thickness, or moisture content. During manufacturing, tools experience wear, vibrations cause changes in machine settings, and electrical fluctuations cause variations in power. Workers may not position parts on fixtures consistently, and physical and emotional stress may affect workers’ consistency. In addition, measurement gauges and human inspection capabilities are not uniform, resulting in measurement error. Similar phenomena occur in service processes because of variation in employee and customer behavior, application of technology, and so on. Reducing variation results in more consistency in manufacturing and service processes, fewer errors, happier customers, and better accuracy of such things as delivery time quotes. Although variation exists everywhere, many managers often do not recognize it or consider it in their decisions. How often do managers make decisions based on one or two data points without looking at the pattern of variation, see trends in data that aren’t justified, or try to manipulate measures they cannot truly control? Unfortunately, the answer is quite often. For example, if sales in some region fell from the previous quarter, a regional manager might quickly blame her sales staff for not working hard enough, even though the drop in sales may simply be the result of uncontrollable variation. Usually, it is simply a matter of ignorance of how to deal with variation in data. This is where business analytics can play a significant role. Statistical analysis can provide better insight into the facts and nature of relationships among the many factors that may have contributed to an event and enable managers to make better decisions. 8
Galen Britz, Don Emerling, Lynne Hare, Roger Hoerl, and Janice Shade, “How to Teach Others to Apply Statistical Thinking,” Quality Progress (June 1997): 67–79.
188
Chapter 4 Descriptive Statistics
EXAMPLE 4.37
Applying Statistical Thinking
Figure 4.50 shows a portion of the data in the Excel file Surgery Infections that document the number of infections that occurred after surgeries over 36 months at one hospital, along with a line chart of the number of infections. (We will assume that the number of surgeries performed each month was the same.) The number of infections tripled in months 2 and 3 as compared to the first month. Is this indicative of a trend caused by failure of some health care protocol or simply random variation? Should action be taken to determine a cause? From a statistical perspective, three points are insufficient to conclude that a trend exists. It is more appropriate to look at a larger sample of data and study the pattern of variation. Over the 36 months, the data clearly indicate that variation exists in the monthly infection rates. The number of infections seems to fluctuate between 0 and 3 with the exception of month 12. However, a visual analysis of the chart cannot necessarily lead to a valid conclusion. So let’s apply some statistical thinking. The average number of infections is 1.583, and the standard deviation is 1.180.
If we apply the empirical rule that most observations should fall within three standard deviations of the mean, we would expect them to fall between  1.957 (clearly, the number of infections cannot be negative, so let’s set this value to zero) and 5.124. This means that, from a statistical perspective, we can expect almost all the observations to fall within these limits. Figure 4.51 shows the chart displaying these ranges. The number of infections for month 12 clearly exceeds the upper range value and suggests that the number of infections for this month is statistically different from the rest. The hospital administrator should seek to investigate what may have happened that month and try to prevent similar occurrences. Similar analyses are used routinely in quality control and other business applications to monitor performance statistically. The proper analytical calculations depend on the type of measure and other factors and are explained fully in books dedicated to quality control and quality management.
Variability in Samples Because we usually deal with sample data in business analytics applications, it is extremely important to understand that different samples from any population will vary; that is, they will have different means, standard deviations, and other statistical measures and will have differences in the shapes of histograms. In particular, samples are extremely sensitive to the sample size—the number of observations included in the samples. ▶▶ Figure
4.50
Portion of Excel File Surgery Infections
Chapter 4 Descriptive Statistics
▶▶ Figure
189
4.51
Infections with Empirical Rule Ranges
EXAMPL E 4.38
Variation in Sample Data
For the 250 observations in the Computer Repair Times database, we can determine that the average repair time is 14.9 days, and the variance of the repair times is 35.50. Suppose we selected some smaller samples from these data. Figure 4.52 shows two samples of size 50 randomly selected from the 250 repair times. Observe that the means and variances
▲▲ Figure
differ from each other as well as from the mean and variance of the entire sample shown in Figure 4.24. In addition, the histograms show a slightly different profile. In Figure 4.53, we show the results for two smaller samples of size 25. Here we actually see more variability in both the statistical measures and the histograms as compared with the entire data set.
4.52
Two Samples of Size 50 from Computer Repair Times
This example demonstrates that it is important to understand the variability in sample data and that statistical information drawn from a sample may not accurately represent the population from which it comes. This is one of the most important concepts in applying business analytics. We explore this topic more in Chapter 6.
190
Chapter 4 Descriptive Statistics
▲▲ Figure
4.53
Two Samples of Size 25 from Computer Repair Times
CHECK YOUR UNDERSTANDING 1. What are the principles of statistical thinking? 2. What factors cause variation in sample data? 3. Why is understanding variation in sample data important?
ANALYTICS IN PRACTICE: Applying Statistical Thinking to
Detecting Financial Problems9
Over the past decade, there have been numerous discoveries of management fraud that have led to the downfall of several prominent companies. These companies had been effective in hiding their financial difficulties, and investors and creditors are now seeking ways to identify financial problems before scandals occur. Even with the passage of the SarbanesOxley Act in July 2002, which helped to improve the quality of the data being disclosed to the public, it is still possible to misjudge an organization’s financial strength without analytical evaluation. Several warning signs exist, but there is no systematic and objective way to determine whether a given financial metric, such as a writeoff or insidertrading pattern, is high or unusual.
Researchers have proposed using statistical thinking to detect anomalies. They propose an “anomaly detection score,” which is the difference between a target financial measure and the company’s own past performance or its competitors’ current performance using standard deviations. This technique is a variation of a standardized zscore. Specifically, their approach involves comparing performance to past performance (within analysis) and comparing performance to the performance of the company’s peers over the same period (between analyses). They created two types of exceptional anomaly scores: zbetween (Zb) to address the variation between companies and zwithin (Zw) to address the variation within the company. These measures quantify the number of standard deviations a
(continued )
9
Based on Deniz Senturk, Christina LaComb, Radu Neagu, and Murat Doganaksoy, “Detect Financial Problems With Six Sigma,” Q uality Progress (April 2006): 41–47.
Chapter 4 Descriptive Statistics
company’s financial measure deviates from the average. Using these measures, the researchers applied the technique to 25 case studies. These included several highprofile companies that had been charged with financial statement fraud by the SEC or had admitted accounting errors, causing a restatement of their financials. The method was able to identify anomalies for critical metrics known by experts to be warning signs for financial statement fraud. These warning signs were consistent when compared with expert postmortem commentary on the highprofile fraud cases. More importantly, they signaled anomalous behavior at least six quarters before an SEC investigation announcement with fewer than 5% false negatives and 40% false positives.
langstrup/123RF
KEY TERMS Arithmetic mean (mean) Bimodal Categorical (nominal) data Chebyshev’s theorem Coefficient of kurtosis (CK) Coefficient of skewness (CS) Coefficient of variation (CV) Contingency table Continuous metric Correlation Correlation coefficient (Pearson product moment correlation coefficient) Covariance Crosstabulation Cumulative relative frequency Cumulative relative frequency distribution Data profile (fractile) Descriptive statistics Discrete metric Dispersion Empirical rules Frequency distribution Histogram Interquartile range (IQR or midspread) Interval data kth percentile Kurtosis
Measure Measurement Median Metric Midrange Mode Ogive Ordinal data Outlier Population Process capability index Proportion Quartile Range Ratio data Relative frequency Relative frequency distribution Return to risk Sample Sample correlation coefficient Skewness Standard deviation Standardized value (zscore) Statistic Statistical thinking Statistics Unimodal Variance
191
192
Chapter 4 Descriptive Statistics
CHAPTER 4 TECHNOLOGY HELP Useful Excel Functions COUNTIF(range, cell_reference) Used to count frequencies to construct a frequency distribution for nongrouped data or to find a proportion. PERCENTILE.INC(array, k) Computes the 100 × kth percentile of data in the range specified in the array field, where k is between 0 and 1, inclusive. QUARTILE.INC(array, quart) Computes quartiles, where array specifies the range of the data and quart is a whole number between 1 and 4, designating the desired quartile. AVERAGE(data range) Finds the average, or arithmetic mean. MEDIAN(data range) Finds the median. MODE.SNGL(data range) Finds the mode. MODE.MULT(data range) Finds an array of modes. VAR.S(data range) Computes the sample variance. VAR.P(data range) Computes the population variance. STDEV.S(data range) Computes the sample standard deviation. STDEV.P(data range) Computes the population standard deviation. STANDARDIZE(x, mean, standard_dev) Calculates a standardized value, or zscore, for the value x given the mean and standard deviation. SKEW.P(data range) Finds the coefficient of skewness for population data. SKEW(data range) Finds the coefficient of skewness for sample data. KURT(data range) Computes the “excess kurtosis” for sample data. COVARIANCE.P(array1, array2) Computes the covariance between two sets of data for a population. COVARIANCE.S(array1, array2) Computes the covariance between two sets of data for a sample. CORREL(array1, array2) Computes the correlation coefficient between two sets of data for either a sample or a population. Excel Techniques Histogram tool (Examples 4.5 and 4.6): Click the Data Analysis tools button in the Analysis group under the Data tab in the Excel menu bar and
select Histogram from the list. In the dialog box, specify the Input Range corresponding to the data and check the Labels box if appropriate. Specify a Bin Range (suggested). Check the Chart Output box to display a histogram in addition to the frequency distribution. Check the other optional boxes to display a Pareto distribution or cumulative frequencies. Descriptive Statistics tool (Example 4.28): Each data set to be analyzed must be in a single row or column. Click on Data Analysis in the Analysis group under the Data tab in the Excel menu bar. Select Descriptive Statistics from the list. Check the box Labels in First Row if labels are included in the input range. Choose whether to save the results in the current worksheet or in a new one. For basic summary statistics, check the box Summary statistics. Displaying statistical information in PivotTables (Example 4.32): Create a PivotTable. In the Value Field Settings dialog, choose the average, standard deviation, or variance of a value field. Correlation tool (Example 4.35): Select Correlation from the Data Analysis tool list. Input the range of the data (which must be in contiguous columns), specify whether the data are grouped by rows or columns, and indicate whether the first row contains data labels. The output provides correlations between every pair of variables. StatCrunch StatCrunch provides many tools for computing and visualizing statistical data. You can find video tutorials and stepbystep procedures with examples at https:// www.statcrunch.com/5.0/example.php. We suggest that you first view the tutorials Getting started with StatCrunch and Working with StatCrunch sessions. The following tutorials, which are located under the Graphs, Summary Statistics and Tables, and Regression and Correlation groups on this Web page, explain how to create frequency distributions, histograms, summary statistics, and correlation tables: ■■ ■■ ■■
Frequency tables Histograms Simple bar plots with raw data
Chapter 4 Descriptive Statistics
■■ ■■ ■■ ■■
Simple bar plots with summary data Summary statistics for columns Summary statistics for rows Correlations between columns
Example: Create a Bar Plot 1. Choose the With data option to use data consisting of individual outcomes in the data table. a. Select the column(s) to be displayed.
193
b. Enter an optional Where statement to specify the data rows to be included. c. Select an optional Group by column to do a sidebyside bar plot. 2. Choose the With summary option to use summary information consisting of categories and counts. a. Select the column containing the categories. b. Select the column containing the counts. 3. Click Compute! to construct the bar plot(s).
PROBLEMS AND EXERCISES Metrics and Data Classification 1. A survey handed out to students at a local university
in Munich, Germany, asked the following: ■■ ■■ ■■ ■■ ■■ ■■
Where do you live? (1 On Campus 2 City 3 Other) How do you feel today? (using a scale of 1–5, ranging from very happy to very unhappy) What is your family’s average annual income? What is your weight in kilograms? (less than 50 kgs, 51–70 kgs, 71–90 kgs, and more than 90 kgs) Length of residency Overall satisfaction with university facilities (using a scale of 1–5, ranging from poor to excellent)
What types of data (categorical, ordinal, interval, or ratio) do each of the survey items represent? Why? 2. Classify each of the data elements in the Sales Trans
actions database as categorical, ordinal, interval, or ratio data and explain why. 3. Identify each of the variables in the Excel file Credit
Approval Decisions as categorical, ordinal, interval, or ratio and explain why. 4. Identify each of the variables in the Excel file Corpo
rate Default Database as categorical, ordinal, interval, or ratio and explain why. 5. Classify each of the variables in the Excel file Wed
dings as categorical, ordinal, interval, or ratio and explain why. Frequency Distributions and Histograms 6. Use the COUNTIF function to construct a frequency
distribution of the types of loans in the Excel file Credit Risk Data, develop a column chart to express the results visually, and compute the relative frequencies.
7. Use the COUNTIF function to construct frequency
distributions for gender, preferred genre, and influence, develop column charts to express the results visually, and compute the relative frequencies based on the data in the Excel file Video Streaming Services. 8. A community health status survey obtained the
following demographic information from the respondents: Age
Frequency
18 to 29
250
30 to 45
740
46 to 64
560
65 to 80
370
Compute the relative frequencies and cumulative relative frequencies of the age groups. 9. Use the Histogram tool to construct frequency distri
butions and histograms for weekly usage and waiting time in the Excel file Car Sharing Report. Do not group the data into bins. 10. Construct frequency distributions and histograms
using the Excel Histogram tool for the gross sales and gross profit data in the Excel file Sales Data. First let Excel automatically determine the number of bins and bin ranges. Then determine a more appropriate set of bins and rerun the Histogram tool. 11. Use the Histogram tool to develop a frequency dis
tribution and histogram for the number of months as a customer of the bank in the Excel file Credit Risk Data. Use your judgment for determining the number of bins to use. Compute the relative and cumulative relative frequencies and use a line chart to construct an ogive.
194
Chapter 4 Descriptive Statistics
12. Identify the bins for number of letters, width, and
23. Use PivotTables to construct crosstabulations
height in the provided data in the Excel file Letter. Then use the Histogram tool to compute frequency distributions and cumulative relative frequencies and create charts for the data.
between each pair of variables in the Excel file Beverage Preferences.
13. Use a PivotTable to construct a frequency distribu
tion and histogram of lunch sales amounts in the Restaurant Sales database. 14. Use a PivotTable to develop a frequency distribution
with six bins for the age of individuals in the Base Data worksheet in the Credit Risk Data file. Compute and chart the relative and cumulative relative frequencies. 15. Use a PivotTable to construct a frequency distribu
tion and histogram for GPA in the Excel file Grade Point Averages. Percentiles and Quartiles 16. Find the 25th and 70th percentiles of the costs in the
Excel file Road Construction. Use formula (4.3), the Excel PERCENTILE.INC function, and the Rank and Percentile tool, and compare the results. 17. Find the 10th and 90th percentiles and the first and
third quartiles for the time difference between the scheduled and actual arrival times in the World Airline Data Excel file. Use formula (4.3), the Excel PERCENTILE.INC function, and the Rank and Percentile tool, and compare the results. 18. Find the first, second, and third quartiles for the com
bined amounts of checking and savings accounts in the Excel file Credit Risk Data and interpret the results. 19. Find the first, second, and third quartiles for the sales
amounts in the Sales Transactions database and interpret the results. CrossTabulations 20. Use a PivotTable to construct a crosstabulation for
loan purpose and credit risk for the Base Data worksheet in the Excel file Credit Risk Data. 21. Use a PivotTable to construct a crosstabulation for
marital status and housing type for the Base Data worksheet in the Excel file Credit Risk Data. 22. Use PivotTables to construct crosstabulations for
(1) neighborhood and improvement and (2) neighborhood and structure home for the Excel file Home.
Descriptive Statistical Measures 24. Find the mean, median, and midrange for the data in
the Excel file Automobile Quality using the appropriate Excel functions or formulas. 25. In the Excel file Facebook Survey, find the mean,
median, and midrange for hours online/week and number of friends in the sample using the appropriate Excel function or formulas. Compare these measures of location. 26. Compute the mean, median, midrange, and mode
for each of the importance factors in the Excel file Coffee Shop Preferences using the appropriate Excel functions. 27. Considering the data in the Excel file House Market
Value as a population of homeowners on this street, compute the mean, variance, and standard deviation for each of the variables using a spreadsheet and formulas (4.4), (4.7), and (4.9). Verify your calculations using the appropriate Excel function. 28. Use the Excel file Course Marks to calculate the
mean, variance and standard deviation for each variable using formulas (4.5), (4.8), and (4.10). Verify your calculations using the appropriate Excel function. 29. In the Excel file Facebook Survey, find the range,
variance, standard deviation, and interquartile range for hours online/week and number of friends in the sample using the appropriate Excel functions. Compare these measures of dispersion. 30. Use the Descriptive Statistics tool to summarize the
mean, median, variance, and standard deviation of the prices of shares in the Coffee Shares Excel file. 31. Use the Excel file Land to find the mean, standard
deviation, and interquartile range of the land size. Show that Chebyshev’s theorem holds for the data and determine how accurate the empirical rules are. 32. The Excel file Atlanta Airline Data provides arrival
and taxiin time statistics for one day at Atlanta Hartsfield International Airport. Find the mean and standard deviation of the difference between the scheduled and actual arrival times and the taxiin
Chapter 4 Descriptive Statistics
time to the gate. Compute the zscores for each of these variables. 33. Compute the mean and standard deviation of the data
in the Cost of Living Adjustments Excel file. Then compute the zscores for the comparative salaries and housing adjustments and interpret your results. 34. Compute the coefficient of variation for each variable
in the Excel file Course Marks. Which has the least and greatest relative dispersion?
for weekly usage and waiting time in the Excel file Car Sharing Report without grouping the data into bins. Use your results and formulas (4.17) and (4.19) to find the mean and sample variance. Check your results using Excel functions with the original data. 42. The amount of time it took 25 respondents to solve
a particular task has been recorded in the following table:
35. Find 30 days of stock prices for three companies in
Time
different industries. The average stock prices should have a wide range of values. Using the data, compute and interpret the coefficient of variation. 36. Apply the Descriptive Statistics tool for subsets of
liberal arts colleges and research universities in the Excel file Colleges and Universities. Compare the two types of colleges. What can you conclude? 37. Use the Descriptive Statistics tool to summarize the
percent gross profit, gross sales, and gross profit in the Excel file Sales Data. 38. Use the Descriptive Statistics tool to summarize the
responses in the Excel file Job Satisfaction. What information can you conclude from this analysis? 39. The Data worksheet in the Excel file Airport Service
Times lists a large sample of the time in seconds to process customers at a ticket counter. The second worksheet shows a frequency distribution and histogram of the data. a. Summarize the data using the Descriptive Statistics tool. What can you say about the shape of the distribution of times? b. Find the 90th percentile. c. How might the airline use these results to manage its ticketing counter operations?
195
Frequency (f)
5 … t 6 10
1
10 … t 6 15
4
15 … t 6 20
6
20 … t 6 25
4
Develop a table to estimate the sample mean and sample standard deviation of the time using formulas (4.17) and (4.19). 43. A teacher recorded the amount of time taken by 21
students to complete a sprint race, to the nearest second, as follows: ■■ ■■ ■■ ■■
Two students: between 51 and less than 56 seconds Seven students: between 56 and less than 61 seconds Eight students: between 61 and less than 66 seconds Four students: between 66 and less than 71 seconds
Develop a table to estimate the sample mean and sample standard deviation for the recorded time using formulas (4.21) and (4.23). 44. The data in the Excel file Church Contributions were
reported on annual giving for a church. Estimate the mean and standard deviation of the annual contributions by implementing formulas (4.20) and (4.22) on a spreadsheet, assuming these data represent the entire population of parishioners.
omputing Descriptive Statistics for C Frequency Distributions
escriptive Statistics for Categorical Data: D The Proportion
40. Construct a frequency distribution for education in the
45. The Excel file EEO Employment Report shows the
Excel file Voter. Develop a table to estimate the sample mean and sample variance using formulas (4.17) and (4.19). Check your results using Excel functions with the original data.
number of people employed in different professions for various racial and ethnic groups. For all employees and each racial/ethnic group, find the proportions of men and women in each profession.
41. In Problem 9, we asked you to use the Histogram tool
46. In the Excel file Bicycle Inventory, find the proportion
to construct frequency distributions and histograms
of bicycle models that sell for less than $200.
196
Chapter 4 Descriptive Statistics
47. In the Excel file Postal Services, find the proportion
tion of customers who used PayPal, the proportion of customers who used credit cards, the proportion that purchased a book, and the proportion that purchased a DVD.
former employees at one of the bank’s call centers (see the Excel file Bank Call Centers). Use PivotTa bles to find these statistics: a. The average length of service for males and females in the sample b. The average length of service for individuals with and without a college degree c. The average length of service for males and females with and without prior call center experience
49. In the Excel file Sales Record, find the proportions
58. For the Peoples Choice Bank database, use PivotTa
of the number of items delivered to Portland, New York, and Chicago for six days in four consecutive weeks. 48. In the Sales Transactions database, find the propor
for the first five variables. Statistics in PivotTables 50. Create a PivotTable to find the mean and standard
deviation of the amount of travel expenses for each sales representative in the Excel file Travel Expenses. 51. Create a PivotTable for the data in the Excel file Wed
ding Services to analyze the average wedding cost by type of pay or and value rating. What conclusions do you reach? 52. The Excel file Sample Sales provides sample data on
the sales representatives’ record. Create a PivotTable to find the following: a. Crosstabulation of region versus item. b. Mean units, unit cost, and accumulative sales in each region by item. State the conclusions for your results. 53. Use a PivotTable to find the mean annual income by
level of education for the data in the Excel file Education and Income. 54. In the Excel file Debt and Retirement Savings, use a
PivotTable to find the mean and standard deviation of income, longterm debt, and retirement savings for both single and married individuals. 55. Using PivotTables, find the mean and standard devia
tion of sales by region in the Sales Transactions database. 56. Use the Excel file Graduation Employment to analyze
the average salary and the average percentage of failure rate in interviews among the graduates from the different colleges. What conclusions do you reach? 57. Call centers have high turnover rates because of
the stressful environment. The national average is approximately 50%. The director of human resources for a large bank has compiled data from about 70
bles to find the average transaction amount for each account and branch. 59. A national homebuilder builds singlefamily homes
and condominiumstyle townhouses. The Excel file House Sales provides information on the selling price, lot cost, type of home, and region of the country for closings during one month. Use PivotTables to find the average selling price and lot cost for each type of home in each region of the market. What conclusions might you reach from this information? Measures of Association 60. Find the correlation between weekly food expendi
ture and monthly household income in the Excel file Food Expenditure. 61. Use the Excel file Test Scores and Sales, developed
by nine salespersons during the past year, to find the coefficient of correlation between the test scores and sales. 62. The Excel file House Sales lists a sample of the area
(in sq. ft) and prices (in dollars) of various houses. Compute the covariance and correlation between the area and prices. 63. For the President’s Inn Guest Data, find the mean
length of stay and mean number of guests per party. Is there any correlation between the size of the party and the length of stay? 64. For the Excel file TV Viewing Survey, is there a sig
nificant correlation between (1) the number of TVs in a home and the hours of viewing per week, and (2) age and hours of viewing per week? 65. For the Excel file Credit Risk Data, compute the cor
relation between age and months employed, age and combined checking and savings account balance, and the number of months as a customer and amount of money in the bank. Interpret your results.
Chapter 4 Descriptive Statistics
Outliers 66. Suppose that the average height of all professional
basketball players is 6 feet with standard deviation of 5 feet. A player of interest to you is 5.5 feet. Would you consider this player as the outlier? Why? 67. Examine the zscores you computed in Problem 32
for the Atlanta Airline Data. Do they suggest any outliers in the data? 68. In the Excel file Weddings Services, find the mean,
median, and standard deviation of the wedding costs. What would you tell a newly engaged couple about what cost to expect? Consider the effect of possible outliers in the data. Using Descriptive Statistics to Analyze Survey Data 69. Use the following approaches to analyze the survey
data in the Excel file Insurance Survey. ■■ ■■ ■■
■■ ■■ ■■
Frequency distributions and histograms for the ratio variables (age, years employed, and satisfaction) Descriptive statistical measures for the ratio variables using the Descriptive Statistics tool Proportions for various attributes of the categorical variables in the sample, such as the proportion of males and females, different educational levels, and marital status PivotTables that break down the averages of ratio variables by gender, education, and marital status Crosstabulations by gender and education, gender and marital status, and education and marital status Zscores for examination of potential outliers
70. The Excel file Auto Survey contains a sample of data
about vehicles owned, whether they were purchased new or used, and other types of data. Use the appropriate statistical tools to analyze these data. Summarize the observations that you can make from these results. 71. A producer of computeraided design software for
the aerospace industry receives numerous calls for technical support. Tracking software is used to monitor response and resolution times. In addition, the company surveys customers who request support using the following scale: 0—Did not meet expectations 1—Marginally met expectations 2—Met expectations
197
3—Exceeded expectations 4—Greatly exceeded expectations The questions are as follows: Q1: Did the support representative explain the
process for resolving your problem? Q2: Did the support representative keep you in
Q3: Q4: Q5: Q6:
formed about the status of progress in resolving your problem? Was the support representative courteous and professional? Was your problem resolved? Was your problem resolved in an acceptable amount of time? Overall, how did you find the service provided by our technical support department?
A final question asks the customer to rate the overall quality of the product using a scale: 0—very poor, 1—poor, 2—good, 3—very good, 4—excellent. A sample of survey responses and associated resolution and response data are provided in the Excel file Customer Support Survey. Use whatever Excel tools you deem appropriate to analyze these sample data and write a report to the manager explaining your findings and conclusions. 72. The Excel file Employee Selection contains role
of employees in an organization, along with other specifics—gender, salary, years of education, date of birth, and whether they are citizens of the country. Analyze the data using all the tools that have been discussed in this chapter. Write a brief report on your analysis. Statistical Thinking in Business Decisions 73. Use the Manufacturing Measurements Data Excel file
to compute sample means, assuming that each row in the data file represents a sample from the manufacturing process. Plot the sample means on a line chart, compute the standard deviation of the sample means (not the individual observations!), add empirical rule ranges for the sample means, and interpret your results. 74. Airbnb is an online marketplace for arranging or
offering lodging, primarily homestays, or tourism experiences. Airbnb provides a platform for hosts to accommodate guests with shortterm lodging and tourismrelated activities. Guests can search for lodging using filters such as lodging type, dates, location,
198
Chapter 4 Descriptive Statistics
and price. However, there are also many controversies around it. For example, Airbnb received criticism for allegedly resulting in increased housing prices. The data in the Excel file Airbnb Samples contains randomly selected samples that include prices for different Airbnb accommodations (same quality, same length of stay). Compute the sample means and the empirical rule ranges. Interpret your findings.
75. Find the mean and variance of a deck of 52 cards,
where an ace is counted as 11 and a picture card as 10. Construct a frequency distribution and histogram of the card values. Shuffle the deck and deal two samples of 20 cards (starting with a full deck each time); compute the mean and variance and construct a histogram. How does the sample data differ from the population data? Repeat this experiment for two samples of 5 cards and summarize your conclusions.
CASE: DROUT ADVERTISING RESEARCH PROJECT The background for this case was introduced in Chapter 2. For this part of the case, summarize the numerical data using frequency distributions and histograms, crosstabulations, PivotTables, and descriptive statistics measures;
find proportions for categorical variables; examine correlations; and so on. Write up your findings in a formal document, or add your findings to the report you completed for the case in Chapter 2 at the discretion of your instructor.
CASE: PERFORMANCE LAWN EQUIPMENT Elizabeth Burke has received several questions from other PLE managers regarding quality, customer satisfaction, and operational performance. She would like you to summarize some data from the Performance Lawn Equipment Database using statistical tools and analysis: 1. Frequency distributions and histograms for the data in the Customer Survey worksheet 2. Descriptive statistical measures for engine production time in the worksheet Engines 3. A frequency distribution and histogram for the blade weight samples in the worksheet Blade Weight
4. The proportion of samples that failed the functional performance test in the worksheet Mower Test 5. PivotTables that summarize the data and provide useful insights in the worksheet Employee Retention 6. Correlations among the satisfaction survey variables in the worksheet Purchasing Survey Write a report that clearly shows these results and explain the key insights.
APPENDIX
Additional Charts for Descriptive Statistics in Excel for Windows
A4
Excel 2016 for Windows provides two additional charts for descriptive statistics that are not available on the Mac platform. The first is the Histogram chart, which can be chosen from the Insert 7 Chart option. You may customize the number of bins and bin widths by right clicking the xaxis and choosing Format Axis from the popup menu. Figure A4.1 shows an example for the cost per order in the Purchase Orders database. The histogram is patterned after Figure 4.11. Select the bin width to be 26,000, and choose the underflow bin (which is the first group) to be the upper group limit of the first bin (26,000), and the overflow bin to be the upper group limit of the last bin (130,000). One useful feature of this chart is that you can easily change the number of bins to create the most visually appealing histogram. The second chart is called a boxandwhisker chart. A boxandwhisker chart displays five key statistics of the data set: the minimum value, first quartile, median, third
▶▶ Figure
A4.1
Histogram of Cost per Order for Purchase Orders Database 199
200
Appendix A4 Additional Charts for Descriptive Statistics in Excel for Windows
quartile, and maximum value. Figure A4.2 shows a boxandwhisker chart for the sales data in the Monthly Product Sales database. The “whiskers” extend on either side of the box to show the minimum and maximum values in the data set. The box encloses the first and third quartiles (that is, the interquartile range, IQR), and the line inside the box shows the median value. Very long whiskers suggest possible outliers in the data. Boxandwhiskers charts show the relative dispersion in the data and the general shape of the distribution (see the discussion of skewness in the chapter). For example, the distribution of product E is relatively symmetric, while that of product D shows more skewness.
▲▲ Figure
A4.2
BoxandWhisker Chart
PROBLEMS AND EXERCISES 1. Use the Histogram chart to construct histograms for
weekly usage and waiting time in the Excel file Car Sharing Survey. 2. Construct histograms using the Histogram chart for
the gross sales and gross profit data in the Excel file Sales Data. 3. For the Base Data in the Excel file Credit Risk Data,
use the Histogram chart to develop a histogram
with a bin width of 5 for the number of months as a customer of the bank. 4. Create boxandwhisker charts for each type of
expense in the Excel file Budget Forecasting. 5. Create boxandwhisker charts for each type of cost
of living adjustment in the Excel file Cost of Living Adjustments.
CHAPTER
Probability Distributions and Data Modeling
5
PeterVrabel/Shutterstock
LEARNING OBJECTIVES ■■
■■
After studying this chapter, you will be able to:
Explain the concept of probability and provide examples of the three definitional perspectives of probability. Use probability rules and formulas to perform probability calculations.
■■
■■
Explain conditional probability and how it can be applied in a business context. Compute conditional probabilities from crosstabulation data.
202 ■■
■■ ■■
■■ ■■ ■■
■■
■■
■■
Chapter 5 Probability Distributions and Data Modeling
Determine if two events are independent using probability arguments. Apply the multiplication law of probability. Explain the difference between a discrete and a continuous random variable. Define a probability distribution. Verify the properties of a probability mass function. Use the cumulative distribution function to compute probabilities over intervals. Compute the expected value and variance of a discrete random variable. Use expected values to support simple business decisions. Calculate probabilities for the Bernoulli, binomial, and Poisson distributions, using the probability mass function and Excel functions.
■■
■■ ■■
■■
■■
■■
■■
■■
Explain how a probability density function differs from a probability mass function. List the key properties of probability density functions. Use the probability density and cumulative distribution functions to calculate probabilities for a uniform distribution. Describe the normal and standard normal distributions and use Excel functions to calculate probabilities. Use the standard normal distribution table and zvalues to compute normal probabilities. Describe properties of the exponential and triangular distributions and compute probabilities. Explain the concepts of distribution fitting and data modeling. Apply the Chi Square Goodness of Fit test.
Most business decisions involve some elements of uncertainty and randomness. For example, the times to repair computers in the Computer Repair Times Excel file that we discussed in Chapter 4 showed quite a bit of uncertainty that we needed to understand to provide information to customers about their computer repairs. We also saw that different samples of repair times result in different means, variances, and frequency distributions. Therefore, it would be beneficial to be able to identify some general characteristics of repair times that would apply to the entire population—even those repairs that have not yet taken place. In other situations, we may not have any data for analysis and simply need to make some judgmental assumptions about future uncertainties. For example, to develop a model to predict the profitability of a new and innovative product, we would need to make reliable assumptions about sales and consumer behavior without any prior data on which to base them. Characterizing the nature of distributions of data and specifying uncertain assumptions in decision models relies on fundamental knowledge of probability concepts and probability distributions—the subject of this chapter. Understanding probability and probability distributions is important in all areas of business. For example, marketing analysts who wish to predict future sales might use probability to assess the likelihood that consumers will purchase their products. In finance, probability is useful in assessing risk of capital investments or product development initiatives. Operations managers apply it routinely in quality control, inventory management, design for product reliability, and customer service policies. Company executives use probability to make
Chapter 5 Probability Distributions and Data Modeling
203
competitive decisions and analyze longterm strategies. Coaches and managers of sports teams use probability to make tactical decisions, such as the best football play to make when faced with third down and 4 yards to go on the opponent’s 17yard line. You use probability concepts in daily life more than you probably realize, such as when deciding whether or not to play golf or go to the beach based on the weather forecast. To experience using probability to make decisions in a game setting, check out one of the author’s favorite games, Qwixx!
Basic Concepts of Probability The notion of probability is used everywhere, both in business and in our daily lives, from market research and stock market predictions to the World Series of Poker and weather forecasts. In business, managers need to know such things as the likelihood that a new product will be profitable or the chances that a project will be completed on time. Probability quantifies the uncertainty that we encounter all around us and is an important building block for business analytics applications. Probability is the likelihood that an outcome— such as whether a new product will be profitable or not or whether a project will be completed within 15 weeks—occurs. Probabilities are expressed as values between 0 and 1, although many people convert them to percentages. The statement that there is a 10% chance that oil prices will rise next quarter is another way of stating that the probability of a rise in oil prices is 0.1. The closer the probability is to 1, the more likely it is that the outcome will occur.
Experiments and Sample Spaces To formally discuss probability, we need some new terminology. An experiment is a process that results in an outcome. An experiment might be as simple as rolling two dice, observing and recording weather conditions, conducting a market research study, or watching the stock market. The outcome of an experiment is a result that we observe; it might be the sum of two dice, a description of the weather, the proportion of consumers who favor a new product, or the change in the Dow Jones Industrial Average (DJIA) at the end of a week. The collection of all possible outcomes of an experiment is called the sample space. For instance, if we roll two fair dice, the possible outcomes for the sum are the numbers 2 through 12; if we observe the weather, the outcome might be clear, partly cloudy, or cloudy; the outcomes for customer reaction to a new product in a market research study would be favorable or unfavorable, and the weekly change in the DJIA can theoretically be any real number (positive, negative, or zero). Note that a sample space may consist of a small number of discrete outcomes or an infinite number of outcomes.
Combinations and Permutations Enumerating and counting the outcomes for an experiment can sometimes be difficult, particularly when the experiment consists of multiple steps. A bit of logic and visualization often helps.
204
Chapter 5 Probability Distributions and Data Modeling
EXAMPLE 5.1
Rolling Two Dice
Suppose we roll two dice. The first roll can be 1, 2, 3, 4, 5, or 6. For each of these outcomes, the second roll can also be 1, 2, 3, 4, 5, or 6. Thus, the outcomes of the experiment are (1, 1), (1, 2), (1, 3), . . . , (6, 4), (6, 5), (6, 6). We can visualize this as a tree diagram in Figure 5.1. A tree diagram is a visualization of a multistep experiment. Counting the outcomes, we find there are 36.
In general, for an experiment with k steps, the number of outcomes is n1 * n2 * c * nk
(5.1)
where ni is the number of possible outcomes in step i. Applying this rule to the dice rolls, we have n1 = 6 and n2 = 6; therefore, the total number of outcomes is 6 * 6 = 36.
In some experiments, we want to select n objects from a set of N objects. EXAMPLE 5.2
Selecting n Objects from N
In a group of five students, an instructor might wish to select three of them to make presentations. How many different ways can this be done? Note that the same student cannot be selected more than once. You might use the analogy of drawing three cards from a deck of five without replacing
▶▶
Figure 5.1
the cards that are drawn. The first student selected can be student 1, 2, 3, 4, or 5. However, if student 1 is selected first, then the second student can be 2, 3, 4, or 5. Then, if the second student selected is student 4, the third student can be 2, 3, or 5. (Try to draw a tree diagram for this!)
First Roll
Second Roll
Outcome
Tree Diagram for Rolling Two Dice
1
1
(1, 1)
2
(1, 2)
6 (1, 6) 1
2
2
(2, 1)
(2, 2)
6 (2, 6) 6 1
2
(6, 1)
(6, 2)
6 (6, 6)
Chapter 5 Probability Distributions and Data Modeling
205
Counting the number of outcomes is not as easy as you might think. Your first thought might be to use formula (5.1) for Example 5.2: We can choose five outcomes in the first step, four in the second, and three in the third, which would result in 5 * 4 * 3 = 60. However, many of these are duplicates, for instance, (1, 2, 3), (1, 3, 2), (2, 1, 3), (2, 3, 1), (3, 1, 2), and (3, 2, 1). Since the order does not matter, we only want to count unique outcomes, which we call combinations. The number of combinations for selecting n objects from a set of N is N! N C(n, N) = a b = (5.2) n n!(N  n)!
The notation ! means factorial, and any number x! is computed as x * (x  1) * (x  2)c * 2 * 1, where x is a nonnegative integer. For instance, 4! = 4 * 3 * 2 * 1 = 24. Zero factorial, 0!, is defined to be 1. EXAMPL E 5.3
Applying the Combinations Formula
In Example 5.2, the number of ways of selecting three students from a group of five is 5 C(3, 5) = a b = 3
5! 3!(5  3)!
=
(5)(4)(3)(2)(1) 3(2)(1) * (2)(1)
= 10
If we want to select n objects from N and the order is important, then we call the outcomes permutations. In general, the number of permutations of n objects selected from N is N! N P(n, N) = n! a b = (5.3) n (N  n)!
EXAMPL E 5.4
Applying the Permutations Formula
In Example 5.2, suppose we want to count the number of ways of selecting three students from a group of five where the order is important (for instance, knowing which student presents first, second, and third). Applying formula (5.3), we have P(3, 5) =
5 3! a b = 3
5! (5  3)!
=
(5)(4)(3)(2)(1) (2)(1)
The permutations formula is easier to apply because we don’t have to think of how many outcomes can occur in each step of the process, particularly when the number of steps is large.
= 60
Probability Definitions Probability may be defined from one of three perspectives. First, if the process that generates the outcomes is known, probabilities can be deduced from theoretical arguments; this is the classical definition of probability.
206
Chapter 5 Probability Distributions and Data Modeling
EXAMPLE 5.5
Classical Definition of Probability
Suppose we roll two dice. If we examine the outcomes described in Example 5.1 and add the values of the dice, we may find the probabilities of rolling each value between 2 and 12. The probability of rolling any number is the ratio of the number of ways of rolling that number to the total number of possible outcomes. For instance, the probability of rolling a 2 is 1>36, the probability of rolling a 3 is 2>36 = 1>18, and the probability of rolling a 7 is 6>36 = 1>6. Similarly, if two consumers are asked whether
or not they like a new product, there could be four possible outcomes: 1. 2. 3. 4.
(like, like) (like, dislike) (dislike, like) (dislike, dislike)
If these are assumed to be equally likely, the probability that at least one consumer would respond unfavorably is 3>4.
The second approach to probability, called the relative frequency definition, is based on empirical data. The probability that an outcome will occur is simply the relative frequency associated with that outcome. EXAMPLE 5.6
Relative Frequency Definition of Probability
Using the sample of computer repair times in the Excel file Computer Repair Times, we developed the relative frequency distribution in Chapter 4, shown again in Figure 5.2. We could state that the probability that a computer would be repaired in as little as four days is 0, the probability that
it would be repaired in exactly ten days is 0.076, and so on. In using the relative frequency definition, it is important to understand that as more data become available, the distribution of outcomes and, hence, the probabilities may change.
Finally, the subjective definition of probability is based on judgment and experience, as financial analysts might use in predicting a 75% chance that the DJIA will increase 10% over the next year, or as sports experts might predict at the start of the football season, a 1in5 chance (0.20 probability) of a certain team making it to the Super Bowl. Which definition to use depends on the specific application and the information we have available. We will see various examples that draw upon each of these perspectives.
▶▶
Figure 5.2
Distribution of Computer Repair Times
Chapter 5 Probability Distributions and Data Modeling
207
Probability Rules and Formulas Suppose we label the n outcomes in a sample space as O1, O2, c, On, where Oi represents the ith outcome in the sample space. Let P1Oi 2 be the probability associated with the outcome Oi. Two basic facts govern probability: ■■
The probability associated with any outcome must be between 0 and 1, inclusive, or
0 … P1Oi 2 … 1 for each outcome Oi
(5.4)
P1O1 2 + P1O2 2 + g + P1On 2 = 1
(5.5)
■■
The sum of the probabilities over all possible outcomes must be 1, or
An event is a collection of one or more outcomes from a sample space. Examples of events are rolling a sum of 7 or 11 with two dice, completing a computer repair in between 7 and 14 days, and obtaining a positive weekly change in the DJIA. This leads to the following rule: Rule 1. The probability of any event is the sum of the probabilities of the outcomes that comprise that event.
EXAMPL E 5.7
Computing the Probability of an Event
Consider the event of rolling a 7 or 11 on two dice. The pro6 and the probability of rolling an 11 bability of rolling a 7 is 36 6 8 2 2 + 36 = 36 . is 36 ; thus, the probability of rolling a 7 or 11 is 36 Similarly, the probability of repairing a computer in seven days or less is the sum of the probabilities of the outcomes
O1 = 0, O2 = 1, O3 = 2, O4 = 3, O5 = 4, O6 = 5, O7 = 6, and O8 = 7 days, or 0.004 + 0.008 + 0.020 = 0.032 (note that the probabilities P(O1) = P(O2) = P(O3) = P(O4) = P(O5) = 0; see Figure 5.2).
If A is any event, the complement of A, denoted Ac, consists of all outcomes in the sample space not in A.
EXAMPL E 5.8
Rule 2. The probability of the complement of any event A is P1Ac 2 = 1  P1A2.
Computing the Probability of the Complement of an Event
If A = 57, 116 in the dice example, then Ac = 52, 3, 4, 5, 6, 8, 9, 10, 126. Thus, the probability of roll8 = 28 ing any sum other than a 7 or 11 is P(Ac) = 1  36 36 . If A = 50, 1, 2, 3, 4, 5, 6, 76 in the computer repair example,
Ac = 58, 9, c, 426 and P(Ac) = 1  0.032 = 0.968. This is the probability of completing the repair in more than a week.
The union of two events contains all outcomes that belong to either of the two events. To illustrate this with rolling the sum of dice, let A be the event {7, 11} and B be the event {2, 3, 12}. The union of A and B is the event {2, 3, 7, 11, 12}. If A and B are two events, the probability that some outcome in either A or B (that is, the union of A and B) occurs is denoted as P(A or B). Finding this probability depends on whether the events are mutually exclusive or not. Two events are mutually exclusive if they have no outcomes in common. The events A and B in the dice example are mutually exclusive. When events are mutually exclusive, the following rule applies: Rule 3. If events A and B are mutually exclusive, then P1A or B2 = P1A2 + P1B2.
208
Chapter 5 Probability Distributions and Data Modeling
EXAMPLE 5.9
Computing the Probability of Mutually Exclusive Events
For the dice example, the probability of event A = 57, 116 8 , and the probability of event B = 52, 3, 126 is is P1A2 = 36 4 P1B2 = 36 . Therefore, the probability that either event A or
B occurs, that is, the sum of the dice is 2, 3, 7, 11, or 12, is 8 4 12 36 + 36 = 36 .
If two events are not mutually exclusive, then adding their probabilities would result in doublecounting some outcomes, so an adjustment is necessary. This leads to the following rule: Rule 4. If two events A and B are not mutually exclusive, then P1A or B2 = P1A2 + P1B2  P1A and B2. Here, (A and B) represents the intersection of events A and B—that is, all outcomes belonging to both A and B. EXAMPLE 5.10 Computing the Probability of Non mutually Exclusive Events In the dice example, let us define the events A = 52, 3, 126 and B = 5even number6. Then A and B are not mutually exclusive because the events have the numbers 2 and 12
in common. Thus, the intersection 1A and B2 = 52, 126. Therefore, P(A or B) = P152, 3, 1262 + P1even number) 20 4 2 P(A and B) = 36 + 18 36  36 = 36 .
Joint and Marginal Probability In many applications, more than one event occurs simultaneously, or in statistical terminology, jointly. We will only discuss the simple case of two events. For instance, suppose that a sample of 100 individuals were asked to evaluate their preference for three new proposed energy drinks in a blind taste test. The sample space consists of two types of outcomes corresponding to each individual: gender (F = female or M = male) and brand preference (B1, B2, or B3). We may define a new sample space consisting of the outcomes that reflect the different combinations of outcomes from these two sample spaces. Thus, for any respondent in the blind taste test, we have six possible (mutually exclusive) combinations of outcomes: 1. O1 2. O2 3. O3 4. O4 5. O5 6. O6
= = = = = =
the respondent is female and prefers brand 1 the respondent is female and prefers brand 2 the respondent is female and prefers brand 3 the respondent is male and prefers brand 1 the respondent is male and prefers brand 2 the respondent is male and prefers brand 3
Here, the probability of each of these events is the intersection of the gender and brand preference event. For example, P1O1 2 = P1F and B1 2, P1O2 2 = P1F and B2 2, and so on. The probability of the intersection of two events is called a joint probability. The probability of an event, irrespective of the outcome of the other joint event, is called a marginal probability. Thus, P1F2, P1M2, P1B1 2, P1B2 2, and P1B3 2 would be marginal probabilities.
Chapter 5 Probability Distributions and Data Modeling
209
EXAMPL E 5.11 Applying Probability Rules to Joint Events Figure 5.3 shows a portion of the data file Energy Drink Survey, along with a crosstabulation constructed from a PivotTable. The joint probabilities of gender and brand preference are easily calculated by dividing the number of respondents corresponding to each of the six outcomes listed above O sub 1 through O sub 6 by the total number of respondents, 100. Thus, P1F and B1 2 = P1O1 2 = 9>100 = 0.09, P1F and B2 2 = P1O2 2 = 6>100 = 0.06, and so on. Note that the sum of the probabilities of all these outcomes is 1. We see that the event F (respondent is female) is composed of the outcomes O1, O2, and O3, and therefore P1F2 = P1O1 2 + P1O2 2 + P1O3 2 = 0.37 using rule 1. The complement of this event is M; that is, the respondent is male. Note that P1M2 = 0.63 = 1  P1F2, as reflected by rule 2. The event B1 is composed of the outcomes O1 and O4, and thus, P1B1 2 = P1O1 2 + P1O4 2 = 0.34. Similarly, we find that P1B2 2 = 0.23 and P1B3 2 = 0.43. Events F and M are mutually exclusive, as are events B1, B2, and B3, since a respondent may be only male or
female and prefer exactly one of the three brands. We can use rule 3 to find, for example, P1B1 o r B2 2 = 0.34 + 0.23 = 0.57. Events F and B1, however, are not mutually exclusive because a respondent can both be female and prefer brand 1. Therefore, using rule 4, we have P1F or B1 2 = P1F2 + P1B1 2  P1F and B1 2 = 0.37 + 0.34  0.09 = 0.62. The joint probabilities can easily be computed, as we have seen, by dividing the values in the crosstabulation by the total, 100. Below the PivotTable in Figure 5.3 is a joint probability table, which summarizes these joint probabilities. The marginal probabilities are given in the margins of the joint probability table by summing the rows and columns. Note, for example, that P1F2 = P1F and B1 2 + P1F and B2 2 + P1F and B3 2 = 0.09 + 0.06 + 0.22 = 0.37. Similarly, P1B1 2 = P1F and B1 2 + P1M and B1 2 = 0.09 + 0.25 = 0.34.
This discussion of joint probabilities leads to the following probability rule: Rule 5. If event A is comprised of the outcomes 5A1, A2, c, An 6 and event B is comprised of the outcomes 5B1, B2, c, Bn 6, then P1Ai 2 = P1Ai and B1 2 + P1Ai and B2 2 + g + P1Ai and Bn 2 P1Bi 2 = P1A1 and Bi 2 + P1A2 and Bi 2 + g + P1An and Bi 2
▲▲
Figure 5.3
Portion of Excel File Energy Drink Survey
210
Chapter 5 Probability Distributions and Data Modeling
Conditional Probability Conditional probability is the probability of occurrence of one event A, given that another event B is known to be true or has already occurred. EXAMPLE 5.12 Computing a Conditional Probability in a CrossTabulation We will use the information shown in the energy drink survey example in Figure 5.3 to illustrate how to compute conditional probabilities from a crosstabulation or joint probability table. Suppose that we know that a respondent is male. What is the probability that he prefers brand 1? From the PivotTable, note that there are only 63 males in the group and
of these, 25 prefer brand 1. Therefore, the probability that a male respondent prefers brand 1 is 25 63 . We could have obtained the same result from the joint probability table by dividing the joint probability 0.25 (the probability that the respondent is male and prefers brand 1) by the marginal probability 0.63 (the probability that the respondent is male).
Conditional probabilities are useful in analyzing data in crosstabulations, as well as in other types of applications. Many companies save purchase histories of customers to predict future sales. Conditional probabilities can help to predict future purchases based on past purchases. EXAMPLE 5.13 Conditional Probability in Marketing The Excel file Apple Purchase History presents a hypothetical history of consumer purchases of Apple products, showing the first and second purchase for a sample of 200 customers who have made repeat purchases (see Figure 5.4). The PivotTable in Figure 5.5 shows the count of the type of second purchase given that each product was purchased first. For example, 13 customers purchased iMacs as their first Apple product. Then the conditional ▶▶
Figure 5.4
Portion of Excel File Apple Purchase History
▶▶
Figure 5.5
PivotTable of Purchase Behavior
probability of purchasing an iPad given that the customer 2 = 0.15. Similarly, 74 customfirst purchased an iMac is 13 ers purchased a MacBook as their first purchase; the conditional probability of purchasing an iPhone if a customer first purchased a MacBook is 26 74 = 0.35. By understanding which products are more likely to be purchased by customers who already own other products, companies can better target advertising strategies.
Chapter 5 Probability Distributions and Data Modeling
211
In general, the conditional probability of an event A given that event B is known to have occurred is P1A ∙ B2 =
P1A and B2 (5.6) P1B2
EXAMPL E 5.14 Using the Conditional Probability Formula Using the data from the energy drink survey example, substitute B1 for A and M for B in formula (5.6). This results in the conditional probability of B1 given M: P1B1 ∙ M2 =
P1B1 and M2 P1M2
=
0.25 = 0.397 0.63
Similarly, the probability of preferring brand 1 if the respondent is female is P1B1 ∙ F2 =
P1B1 and F2 P1F2
=
0.09 = 0.243 0.37
The following table summarizes the conditional probabilities of brand preference given gender.
Brand 1
Brand 2
Brand 3
Male
0.397
0.270
0.333
Female
0.243
0.162
0.595
P(Brand ∣ Gender)
Such information can be important in marketing efforts. Knowing that there is a difference in preference by gender can help focus advertising. For example, we see that about 40% of males prefer brand 1, whereas only about 24% of females do, and a higher proportion of females prefer brand 3. This suggests that it would make more sense to focus on advertising brand 1 more in maleoriented media and brand 3 in femaleoriented media.
We read the notation P1A ∙ B2 as “the probability of A given B.” The conditional probability formula may be used in other ways. For example, multiplying both sides of formula (5.6) by P(B), we obtain P1A and B2 = P1A ∙ B2 P1B2. Note that we may switch the roles of A and B and write P1B and A2 = P1B ∙ A2 P1A2. But P(B and A) is the same as P(A and B); thus we can express P(A and B) in two ways:
P1A and B2 = P1A ∙ B2 P1B2 = P1B ∙ A2 P1A2 (5.7)
This is often called the multiplication law of probability. We may use this concept to express the probability of an event in a joint probability table in a different way. Using the energy drink survey in Figure 5.3 again, note that P1F2 = P1F and Brand 12 + P1F and Brand 22 + P1F and Brand 32 Using formula (5.7), we can express the joint probabilities P(A and B) by P1A ∙ B2 P1B2. Therefore, P1F2 = P1F ∙ Brand 12 P1Brand 12 + P1F ∙ Brand 22 P1Brand 22 + P1F ∙ Brand 32 P1Brand 32 = 10.265210.342 + 10.261210.232 + 10.512210.432 = 0.371 within rounding precision2. We can express this calculation using the following extension of the multiplication law of probability. Suppose B1, B2, c, Bn are mutually exclusive events whose union comprises the entire sample space. Then
P1A2 = P1A ∙ B1 2P1B1 2 + P1A ∙ B2 2P1B2 2 + g + P1A ∙ Bn 2P1Bn 2(5.8)
212
Chapter 5 Probability Distributions and Data Modeling
EXAMPLE 5.15 Using the Multiplication Law of Probability Texas Holdem has become a popular game because of the publicity surrounding the World Series of Poker. At the beginning of a game, players each receive two cards face down (we won’t worry about how the rest of the game is played). Suppose that a player receives an ace on her first card. The probability that she will end up with “pocket aces” (two aces in the hand) is P1ace on first card and ace on second card2 = P1ace on second card ∙ ace on first card2 *
P1ace on first card2. Since the probability of an ace on the first card is 4/52 and the probability of an ace on the second card if she has already drawn an ace is 3/51, we have P1ace on first card and ace on second card2 = P1ace on second card ∙ ace on first card2 * P 1ace on first card2 3 4 = a b * a b = 0.004525 51 52
In Example 5.14, we see that the probability of preferring a brand depends on gender. We may say that brand preference and gender are not independent. We may formalize this concept by defining the notion of independent events: Two events A and B are independent if P1A ∙ B2 = P1A2. EXAMPLE 5.16 Determining if Two Events Are Independent We use this definition in the energy drink survey example. Recall that the conditional probabilities of brand preference given gender are
Brand 1
Brand 2
Brand 3
Male
0.397
0.270
0.333
Female
0.243
0.162
0.595
P(Brand ∙ Gender)
We see that whereas P1B1 ∙M2 = 0.397, P1B1 2 was shown to be 0.34 in Example 5.11; thus, these two events are not independent.
Finally, we see that if two events are independent, then we can simplify the multiplication law of probability in equation (5.7) by substituting P(A) for P1A ∙ B2:
P1A and B2 = P1A2P1B2 = P1B2 P1A2 (5.9)
EXAMPLE 5.17 Using the Multiplication Law for Independent Events Suppose A is the event that a sum of 6 is first rolled on a pair of dice and B is the event of rolling a sum of 2, 3, or 12 on the next roll. These events are
independent because the roll of a pair of dice does not depend on the previous roll. Then we may compute 5 20 P1A and B2 = P1A2P1B2 = 1 36 2 1 364 2 = 1296 .
CHECK YOUR UNDERSTANDING 1. 2. 3. 4. 5. 6.
Define the terms experiment, outcome, and sample space. Explain the difference between a permutation and a combination. Give an example of each of the three definitions of probability. What are the two key facts that govern probability? What is an event? Explain how to compute P(A or B) for two events A and B. Explain the concepts of joint, marginal, and conditional probability, and independent events.
Chapter 5 Probability Distributions and Data Modeling
213
Random Variables and Probability Distributions Some experiments naturally have numerical outcomes, such as a sum of dice, the time it takes to repair computers, or the weekly change in a stock market index. For other experiments, such as obtaining consumer response to a new product, the sample space is categorical. To have a consistent mathematical basis for dealing with probability, we would like the outcomes of all experiments to be numerical. A random variable is a numerical description of the outcome of an experiment. Formally, a random variable is a function that assigns a real number to each element of a sample space. If we have categorical outcomes, we can associate an arbitrary numerical value to them. For example, if a consumer likes a product in a market research study, we might assign this outcome a value of 1; if the consumer dislikes the product, we might assign this outcome a value of 0. Random variables are usually denoted by capital italic letters, such as X or Y. Random variables may be discrete or continuous. A discrete random variable is one for which the number of possible outcomes can be counted. A continuous random variable has outcomes over one or more continuous intervals of real numbers. EXAMPL E 5.18 Discrete and Continuous Random Variables The outcomes of the sum of rolling two dice (the numbers 2 through 12) and customer reactions to a product (like or dislike) are discrete random variables. The number of outcomes may be finite or theoretically infinite, such as the number of hits on a Web site link during some period of time—we cannot place a guaranteed upper limit on this
number; nevertheless, the number of hits can be counted. Examples of continuous random variables are the weekly change in the DJIA, the daily temperature, the time to complete a task, the time between failures of a machine, and the return on an investment.
A probability distribution is the characterization of the possible values that a random variable may assume along with the probability of assuming these values. A probability distribution can be either discrete or continuous, depending on the nature of the random variable it models. Discrete distributions are easier to understand and work with, and we deal with them first. We may develop a probability distribution using any one of the three perspectives of probability. First, if we can quantify the probabilities associated with the values of a random variable from theoretical arguments, then we can easily define the probability distribution. EXAMPL E 5.19 Probability Distribution of Dice Rolls The probabilities of the sum of outcomes for rolling two dice are calculated by counting the number of ways to roll each sum divided by the total number of possible
outcomes. These, along with an Excel column chart depicting the probability distribution, are shown from the Excel file Dice Rolls in Figure 5.6.
Second, we can calculate the relative frequencies from a sample of empirical data to develop a probability distribution. Thus, the relative frequency distribution of computer repair times (Figure 5.2) is an example. Because this is based on sample data, we usually call this an empirical probability distribution. An empirical probability distribution is an approximation of the probability distribution of the associated random variable, whereas the probability distribution of a random variable, such as the one derived from counting arguments, is a theoretical model of the random variable.
214 ▶▶
Chapter 5 Probability Distributions and Data Modeling
Figure 5.6
Probability Distribution of Rolls of Two Dice
Finally, we could simply specify a probability distribution using subjective values and expert judgment. This is often done in creating decision models for phenomena for which we have no historical data. EXAMPLE 5.20 A Subjective Probability Distribution Figure 5.7 shows a hypothetical example of the distribution of one expert’s assessment of how the DJIA might change in the next year. This might have been created purely by
intuition and expert judgment, but we hope it would be supported by some extensive analysis of past and current data using business analytics tools.
Researchers have identified many common types of probability distributions that are useful in a variety of applications of business analytics. A working knowledge of common families of probability distributions is important for several reasons. First, it can help you to understand the underlying process that generates sample data. We investigate the relationship between distributions and samples later. Second, many phenomena in business and nature follow some theoretical distribution and, therefore, are useful in building decision models. Finally, working with distributions is essential in computing probabilities of occurrence of outcomes to assess risk and make decisions.
▶▶
Figure 5.7
Subjective Probability Distribution of DJIA Change
Chapter 5 Probability Distributions and Data Modeling
215
CHECK YOUR UNDERSTANDING 1. E xplain the difference between a discrete and a continuous random variable, and give an example of each. 2. What is a probability distribution? 3. What is an empirical probability distribution? 4. Why is it important to understand the common types of probability distributions?
Discrete Probability Distributions For a discrete random variable X, the probability distribution of the discrete outcomes is called a probability mass function and is denoted by a mathematical function, f(x). The symbol xi represents the ith value of the random variable X, and f1xi 2 is the probability associated with xi. EXAMPL E 5.21 Probability Mass Function for Rolling Two Dice 5 = 0.1389 36 6 f1x6 2 = = 0.1667 36 5 f1x7 2 = = 0.1389 36 4 f1x8 2 = = 0.1111 36 3 f1x9 2 = = 0.0833 36 2 f1x10 2 = = 0.0556 36 1 f1x11 2 = = 0.0278 36
For instance, in Figure 5.6 for the dice example, the values of the random variable X, which represents the sum of the rolls of two dice, are x1 = 2, x2 = 3, x3 = 4, x4 = 5, x5 = 6, x6 = 7, x7 = 8, x8 = 9, x9 = 10, x10 = 11, and x11 = 12. The probability mass function for X is 1 36 2 f1x2 2 = 36 3 f1x3 2 = 36 4 f1x4 2 = 36 f1x1 2 =
f1x5 2 =
= 0.0278 = 0.0556 = 0.0833 = 0.1111
A probability mass function has the properties that (1) the probability of each outcome must be between 0 and 1, inclusive, and (2) the sum of all probabilities must add to 1; that is,
0 … f1xi 2 … 1 for all i (5.10)
a f1xi 2 = 1 (5.11) i
You can easily verify that this holds in each of the examples we have described. A cumulative distribution function, F(x), specifies the probability that the random variable X assumes a value less than or equal to a specified value, x. This is also denoted as P1X … x2 and reads as “the probability that the random variable X is less than or equal to x.”
216
Chapter 5 Probability Distributions and Data Modeling
EXAMPLE 5.22 Using the Cumulative Distribution Function The cumulative distribution function for the sum of rolling two dice is shown in Figure 5.8, along with an Excel line chart that describes it visually from the worksheet CumDist in the Dice Rolls Excel file. To use this, suppose we want to know the probability of rolling a 6 or less. We simply look up the cumulative probability for 6, which is 0.4167. Alternatively, we could locate the point for x = 6 in the chart and estimate the probability from the graph. Also note that since the probability of rolling a 6 or less is 0.4167, then the probability of the complementary event (rolling a 7 or more) is 1  0.4167 = 0.5833. We can also use the cumulative
distribution function to find probabilities over intervals. For example, to find the probability of rolling a number between 4 and 8, P14 … X … 82, we can find P1X … 82 and subtract P1X … 32; that is, P14 … X … 82 = P1X … 82  P1X … 32 = 0.7222  0.0833 = 0.6389 A word of caution. Be careful with the endpoints when computing probabilities over intervals for discrete distributions; because 4 is included in the interval we wish to compute, we need to subtract P1X … 32, not P1X … 42.
Expected Value of a Discrete Random Variable The expected value of a random variable corresponds to the notion of the mean, or average, for a sample. For a discrete random variable X, the expected value, denoted E[X], is the weighted average of all possible outcomes, where the weights are the probabilities:
E3X4 = a xi f 1xi 2 (5.12) ∞
i=1
Note the similarity to computing the population mean using formula (4.16) in Chapter 4: a fi xi N
m =
i=1
N If we write this as the sum of xi multiplied by 1fi >N2, then we can think of fi >N as the probability of xi. Then this expression for the mean has the same basic form as the expected value formula.
▶▶
Figure 5.8
Cumulative Distribution Function for Rolling Two Dice
Chapter 5 Probability Distributions and Data Modeling
217
EXAMPL E 5.23 Computing the Expected Value We may apply formula (5.12) to the probability distribution for the sum of rolling two dice. We multiply the outcome 2 by its probability 1/36, add this to the product of the outcome 3 and its probability, and so on. Continuing in this fashion, the expected value is E[X ] = 210.02782 + 310.05562 + 410.08332 + 510.01112 + 610.13892 + 710.16672 + 810.13892 + 910.1112 + 1010.08332 + 1110.05562 + 1210.02782 = 7
Figure 5.9 shows these calculations in an Excel spreadsheet (worksheet Expected Value in the Dice Rolls Excel file). Note that you can use the SUMPRODUCT function to easily calculate the expected value. For this example, we would use =SUMPRODUCT(A4:A14, B4:B14). As expected (no pun intended), the average value of the sum of the roll of two dice is 7.
Using Expected Value in Making Decisions Expected value can be helpful in making a variety of decisions, even those we see in daily life. EXAMPL E 5.24 Expected Value on Television One of the author’s favorite examples stemmed from a task in season 1 of the former TV show, The Apprentice. Teams were required to select an artist and sell his or her art for the highest total amount of money. One team selected a mainstream artist who specialized in abstract art that sold for between $1,000 and $2,000; the second team chose an avantgarde artist whose surrealist and rather controversial art was priced much higher. Guess who won? The first team did, because the probability of selling a piece of mainstream art was much higher than the avantgarde artist whose bizarre art (the team members themselves didn’t even like it!) had a very low probability of a sale. A simple expected value calculation would have easily predicted the winner. A popular game show that took TV audiences by storm several years ago was called Deal or No Deal. The game involved a set of numbered briefcases that contain amounts of money from 1 cent to $1,000,000. Contestants begin choosing cases to be opened and removed, and their amounts are shown. After each set of cases is opened,
the banker offers the contestant an amount of money to quit the game, which the contestant may either choose or reject. Early in the game, the banker’s offers are usually less than the expected value of the remaining cases, providing an incentive to continue. However, as the number of remaining cases becomes small, the banker’s offers approach or may even exceed the average of the remaining cases. Most people press on until the bitter end and often walk away with a smaller amount than they could have had they been able to estimate the expected value of the remaining cases and make a more rational decision. In one case, a contestant had five briefcases left with $100, $400, $1,000, $50,000, and $300,000. Because the choice of each case is equally likely, the expected value was 0.21$100 + $400 + $1000 + $50,000 + $300,0002 = $70,300, and the banker offered $80,000 to quit. Instead, she said “No deal” and proceeded to open the $300,000 suitcase, eliminating it from the game. She took the next banker’s offer of $21,000, which was more than 60% larger than the expected value of the remaining cases.1
It is important to understand that the expected value is a “longrun average” and is appropriate for decisions that occur on a repeated basis. For onetime decisions, however, you need to consider the downside risk and the upside potential of the decision. The following example illustrates this.
1
“Deal or No Deal: A Statistical Deal.” www.pearsonified.com/2006/03/deal_or_no_deal_the_real_deal.php
218 ▶▶
Chapter 5 Probability Distributions and Data Modeling
Figure 5.9
Expected Value Calculations for Rolling Two Dice
EXAMPLE 5.25 Expected Value of a Charitable Raffle Suppose that you are offered the chance to buy one of 1,000 tickets sold in a charity raffle for $50, with the prize being $25,000. Clearly, the probability of winning 1 , or 0.001, whereas the probability of losing is is 1,000 1  0.001 = 0.999. The random variable X is your net winnings, and its probability distribution is
x
f(x)
 $50
0.999
$24,950
0.001
repeatedly over the long run, you would lose an average of $25.00 each time you play. Of course, for any one game, you would either lose $50 or win $24,950. So the question becomes, Is the risk of losing $50 worth the potential of winning $24,950? Although the expected value is negative, you might take the chance because the upside potential is large relative to what you might lose, and, after all, it is for charity. However, if your potential loss is large, you might not take the chance, even if the expected value were positive.
The expected value, E3X4, is  $5010.9992 + $24,950 10.0012 =  $25.00. This means that if you played this game
Decisions based on expected values are common in real estate development, day trading, and pharmaceutical research projects. Drug development is a good example. The cost of research and development projects in the pharmaceutical industry is generally in the hundreds of millions of dollars and often approaches $1 billion. Many projects never make it to clinical trials or might not get approved by the Food and Drug Administration. Statistics indicate that 7 of 10 products fail to return the cost of the company’s capital. However, large firms can absorb such losses because the return from one or two blockbuster drugs can easily offset these losses. On an average basis, drug companies make a net profit from these decisions. EXAMPLE 5.26 Airline Revenue Management Let us consider a simplified version of the typical revenue management process that airlines use. At any date prior to a scheduled flight, airlines must make a decision as to whether to reduce ticket prices to stimulate demand for unfilled seats. If the airline does not discount the fare, empty seats might not be sold and the airline will lose revenue. If the airline discounts the remaining seats too early
(and could have sold them at the higher fare), they will lose profit. The decision depends on the probability p of selling a fullfare ticket if they choose not to discount the price. Because an airline makes hundreds or thousands of such decisions each day, the expected value approach is appropriate.
Chapter 5 Probability Distributions and Data Modeling
Assume that only two fares are available: full and discount. Suppose that a fullfare ticket is $560, the discount fare is $400, and p = 0.75. For simplification, assume that if the price is reduced, then any remaining seats would be sold at that price. The expected value of not discounting the price is 0.25(0) + 0.75($560) = $420. Because this is higher than the discounted price, the airline should not discount at this time. In reality, airlines constantly update the probability p based on the information they collect and analyze in a
219
database. When the value of p drops below the breakeven point: $400 = p 1 $5602, or p = 0.714, then it is beneficial to discount. It can also work in reverse; if demand is such that the probability that a higherfare ticket would be sold, then the price may be adjusted upward. This is why published fares constantly change and why you may receive lastminute discount offers or may pay higher prices if you wait too long to book a reservation. Other industries such as hotels and cruise lines use similar decision strategies.
Variance of a Discrete Random Variable We may compute the variance, Var[X], of a discrete random variable X as a weighted average of the squared deviations from the expected value:
Var [X] = a 1xi  E[X]2 2f1xi 2 (5.13) ∞
i=1
EXAMPL E 5.27 Computing the Variance of a Random Variable We may apply formula (5.13) to calculate the variance of the probability distribution for the sum of rolling two dice.
Figure 5.10 shows these calculations in an Excel spreadsheet (worksheet Variance in Random Variable Calculations Excel file).
Similar to our discussion in Chapter 4, the variance measures the uncertainty of the random variable; the higher the variance, the higher the uncertainty of the outcome. Although variances are easier to work with mathematically, we usually measure the variability of a random variable by its standard deviation, which is simply the square root of the variance.
Bernoulli Distribution
▶▶
Figure 5.10
Variance Calculations for Rolling Two Dice
The Bernoulli distribution characterizes a random variable having two possible outcomes, each with a constant probability of occurrence. Typically, these outcomes represent “success” 1x = 12, having probability p, and “failure” 1x = 02, having probability
220
Chapter 5 Probability Distributions and Data Modeling
1  p. A success can be any outcome you define. For example, in attempting to boot a new computer just off the assembly line, we might define a success as “does not boot up” in defining a Bernoulli random variable to characterize the probability distribution of a defective product. Thus, success need not be a favorable result in the traditional sense. The probability mass function of the Bernoulli distribution is f1x2 = e
p 1  p
if x = 1 (5.14) if x = 0
where p represents the probability of success. The expected value is p, and the variance is p11  p2. EXAMPLE 5.28 Using the Bernoulli Distribution A Bernoulli distribution might be used to model whether an individual responds positively ( x = 1) or negatively ( x = 0) to a telemarketing promotion. For example, if you estimate that 20% of customers contacted will make a purchase, the probability distribution that describes whether or not a particular individual makes a purchase is Bernoulli with
p = 0.2. Think of the following experiment. Suppose that you have a box with 100 marbles, 20 red and 80 white. For each customer, select one marble at random (and then replace it). The outcome will have a Bernoulli distribution. If a red marble is chosen, then that customer makes a purchase; if it is white, the customer does not make a purchase.
Binomial Distribution The binomial distribution models n independent repetitions of a Bernoulli experiment, each with a probability p of success. The random variable X represents the number of successes in these n experiments. In the telemarketing example, suppose that we call n = 10 customers, each of which has a probability p = 0.2 of making a purchase. Then the probability distribution of the number of positive responses obtained from ten customers is binomial. Using the binomial distribution, we can calculate the probability that exactly x customers out of the ten will make a purchase for any value of x between 0 and 10, inclusive. A binomial distribution might also be used to model the results of sampling inspection in a production operation or the effects of drug research on a sample of patients. The probability mass function for the binomial distribution is
f1x2 = d
n a b px 11  p2 n  x, x
0,
for x = 0, 1, 2, c, n
(5.15)
otherwise
n We saw the notation a b earlier in this chapter when discussing combinations; it repx resents the number of ways of choosing x distinct items from a group of n items and is computed using formula (5.2). EXAMPLE 5.29 Computing Binomial Probabilities We may use formula (5.15) to compute binomial probabilities. For example, if the probability that any individual will make a purchase from a telemarketing solicitation is 0.2, then the probability distribution that x individuals out of ten calls will make a purchase is
f1x2 = d
a
10 b 10.22 x 10.82 10  x, x
0,
for x = 0, 1, 2, c, n otherwise
Chapter 5 Probability Distributions and Data Modeling
Thus, to find the probability that three people will make a purchase among the ten calls, we compute
221
10 f(3) = a 3 b(0.2)3(0.8)10  3
= (10!/(3!7!))(0.008)(0.2097152) = 120(0.008)(0.2097152) = 0.20133
The formula for the probability mass function for the binomial distribution is rather complex, and binomial probabilities are tedious to compute by hand; however, they can easily be computed in Excel using the function BINOM.DIST1number_s, trials, probability_s, cumulative2 In this function, number_s plays the role of x, and probability_s is the same as p. If cumulative is set to TRUE, then this function will provide cumulative probabilities; otherwise the default is FALSE, and it provides values of the probability mass function, f(x). EXAMPL E 5.30 Using Excel’s Binomial Distribution Function Figure 5.11 shows the results of using this function to compute the distribution for the previous example (Excel file Binomial Probabilities). For instance, the probability that exactly three individuals will make a purchase is BINOM.DIST(A10, $B$3, $B$4, FALSE) = 0.20133 = f132.
The probability that three or fewer individuals will make a purchase is BINOM.DIST(A10, $B$3, $B$4, TRUE) = 0.87913 = F132. Correspondingly, the probability that more than three out of ten individuals will make a purchase is 1  F132 = 1  0.87913 = 0.12087.
The expected value of the binomial distribution is np, and the variance is np11  p2. The binomial distribution can assume different shapes and amounts of skewness, depending on the parameters. Figure 5.12 shows an example when p = 0.8. For larger values of p, the binomial distribution is negatively skewed; for smaller values, it is positively skewed. When p = 0.5, the distribution is symmetric.
Poisson Distribution The Poisson distribution is a discrete distribution used to model the number of occurrences in some unit of measure—for example, the number of customers arriving at a Subway store during a weekday lunch hour, the number of failures of a machine during a month, the number of visits to a Web page during 1 minute, or the number of errors per line of software code. ▶▶
Figure 5.11
Computing Binomial Probabilities in Excel
222 ▶▶
Chapter 5 Probability Distributions and Data Modeling
Figure 5.12
Example of the Binomial Distribution with p = 0.8
The Poisson distribution assumes no limit on the number of occurrences (meaning that the random variable X may assume any nonnegative integer value), that occurrences are independent, and that the average number of occurrences per unit is a constant, l (Greek lowercase lambda). The expected value of the Poisson distribution is l, and the variance also is equal to l. The probability mass function for the Poisson distribution is
ellx , f1x2 = c x! 0,
for x = 0, 1, 2, c
(5.16)
otherwise
EXAMPLE 5.31 Computing Poisson Probabilities Suppose that, on average, the number of customers arriving at Subway during lunch hour is 12 customers per hour. The probability that exactly x customers will arrive during the hour is given by a Poisson distribution with a mean of 12. The probability that exactly x customers will arrive during the hour would be calculated using formula (5.16):
e1212x , f(x) = c x! 0,
for x = 0, 1, 2, c otherwise
Substituting x = 5 in this formula, the probability that exactly five customers will arrive is f(5) = 0.01274.
Like the binomial, Poisson probabilities are cumbersome to compute by hand. Probabilities can easily be computed in Excel using the function POISSON.DIST(x, mean, cumulative).
EXAMPLE 5.32 Using Excel’s Poisson Distribution Function Figure 5.13 shows the results of using this function to compute the distribution for Example 5.31 with l = 12 (see the Excel file Poisson Probabilities). Thus, the probability of exactly one arrival during the lunch hour is calculated by the Excel function = POISSON.DIST(A7, $B$3, FALSE) = 0.00007 = f(1); the probability of four arrivals or fewer is calculated by = POISSON.DIST(A10,
$B$3, TRUE) = 0.00760 = F(4), and so on. Because the possible values of a Poisson random variable are infinite, we have not shown the complete distribution. As x gets large, the probabilities become quite small. Like the binomial, the specific shape of the distribution depends on the value of the parameter l; the distribution is more skewed for smaller values.
▶▶
Chapter 5 Probability Distributions and Data Modeling
Figure 5.13
Computing Poisson Probabilities in Excel
ANALYTICS IN PRACTICE: Using the Poisson Distribution for
Priceline is well known for allowing customers to name their own prices (but not the service providers) in bidding for services such as airline flights or hotel stays. Some hotels take advantage of Priceline’s approach to fill empty rooms for leisure travelers while not diluting the business market by offering discount rates through traditional channels. In one study using business analytics to develop a model to optimize pricing strategies for Kimpton Hotels, which develops, owns, or manages more than 40 independent boutique lifestyle hotels in the United States and Canada, the distribution of the number of bids for a given number of days before arrival was modeled as a Poisson distribution because it corresponded well with data that were observed. For example, the average number of bids placed per day three days before arrival on a weekend (the random variable X) was 6.3. Therefore, the distribution used in the model was f(x) = e6.36.3x >x!, where x is the number of bids placed. The analytic model helped to determine the prices to post on Priceline and the inventory allocation for each price. After using the model, rooms sold via Priceline increased 11% in one year, and the average rate for these rooms increased 3.7%. 2
Based on Chris K. Anderson, “Setting Prices on Priceline,” Interfaces, 39, 4 (July–August 2009): 307–315.
Fantasista/Fotolia
Modeling Bids on Priceline2
223
224
Chapter 5 Probability Distributions and Data Modeling
CHECK YOUR UNDERSTANDING 1. What properties must a probability mass function have? 2. W hat is the difference between a probability mass function and a cumulative d istribution function? 3. Explain how to compute the expected value and variance of a discrete random variable. 4. How can expected value concepts be used in business decisions? 5. P rovide examples of situations for which the Bernoulli, binomial, and Poisson distributions can be used.
Continuous Probability Distributions As we noted earlier, a continuous random variable is defined over one or more intervals of real numbers and, therefore, has an infinite number of possible outcomes. Suppose that the expert who predicted the probabilities associated with next year’s change in the DJIA in Figure 5.7 kept refining the estimates over larger and larger ranges of values. Figure 5.14 shows what such a probability distribution might look like using 2.5% increments rather than 5%. Notice that the distribution is similar in shape to the one in Figure 5.7 but simply has more outcomes. If this refinement process continues, then the distribution will approach the shape of a smooth curve, as shown in the figure. Such a curve that characterizes outcomes of a continuous random variable is called a probability density function and is described by a mathematical function f(x).
Properties of Probability Density Functions A probability density function has the following properties: 1. f1x2 Ú 0 for all values of x. This means that a graph of the density function must lie at or above the xaxis. 2. The total area under the density function above the xaxis is 1. This is analogous to the property that the sum of all probabilities of a discrete random variable must add to 1. ▶▶
Figure 5.14
Refined Probability Distribution of DJIA Change
Chapter 5 Probability Distributions and Data Modeling
225
3. P1X = x2 = 0. For continuous random variables, it does not make mathematical sense to attempt to define a probability for a specific value of x because there are an infinite number of values. 4. Probabilities of continuous random variables are only defined over intervals. Thus, we may calculate probabilities between two numbers a and b, P1a … X … b2, or to the left or right of a number c—for example, P1X 6 c2 and P1X 7 c2. 5. P1a … X … b2 is the area under the density function between a and b. The cumulative distribution function for a continuous random variable is denoted the same way as for discrete random variables, F(x), and represents the probability that the random variable X is less than or equal to x, P1X … x2. Intuitively, F(x) represents the area under the density function to the left of x. F(x) can often be derived mathematically from f(x). Knowing F(x) makes it easy to compute probabilities over intervals for continuous distributions. The probability that X is between a and b is equal to the difference of the cumulative distribution function evaluated at these two points; that is,
P1a … X … b2 = P1X … b2  P1X … a2 = F1b2  F1a2 (5.17)
For continuous distributions, we need not be concerned about the endpoints, as we were with discrete distributions, because P1a … X … b2 is the same as P1a 6 X 6 b2 as a result of property 3 above. The formal definitions of expected value and variance for a continuous random variable are similar to those for a discrete random variable; however, to understand them, we must rely on notions of calculus, so we do not discuss them in this book. We simply state them when appropriate.
Uniform Distribution The uniform distribution characterizes a continuous random variable for which all outcomes between some minimum and maximum value are equally likely. The uniform distribution is often assumed in business analytics applications when little is known about a random variable other than reasonable estimates for minimum and maximum values. The parameters a and b are chosen judgmentally to reflect a modeler’s best guess about the range of the random variable. For a uniform distribution with a minimum value a and a maximum value b, the density function is
1 , f 1x2 = c b  a 0,
for a … x … b
(5.18)
otherwise
and the cumulative distribution function is
0, x  a F1x2 = d , b  a 1,
if x 6 a if a … x … b (5.19) if b 6 x
Although Excel does not provide a function to compute uniform probabilities, the formulas are simple enough to incorporate into a spreadsheet. Probabilities are also easy to compute for the uniform distribution because of the simple geometric shape of the density function, as Example 5.33 illustrates.
226
Chapter 5 Probability Distributions and Data Modeling
EXAMPLE 5.33 Computing Uniform Probabilities Suppose that sales revenue, X, for a product varies uniformly each week between a = $1,000 and b = $2,000. The density function is f1x2 = 1> 12,000  1,0002 = 1>1,000 and is shown in Figure 5.15. Note that the area under the density function is 1, which you can easily verify by multiplying the height by the width of the rectangle. Suppose we wish to find the probability that sales revenue will be less than x = $1,300. We could do this in two ways. First, compute the area under the density function using geometry, as shown in Figure 5.16. The area is 11>1,0002 13002 = 0.30. Alternatively, we could use formula (5.19) to compute f(1,300):
Now suppose we wish to find the probability that revenue will be between $1,500 and $1,700. Again, using geometrical arguments (see Figure 5.17), the area of the rectangle between $1,500 and $1,700 is 11>1,0002 12002 = 0.2. We may also use formula (5.17) and compute it as follows: P(1,500 … X … 1,700) = P(X … 1,700)  P(X … 1,500) = F(1,700)F (1,500) =
11,700  1,0002
12,000  1,0002 = 0.7  0.5 = 0.2

(1,500  1,000) (2,000  1,000)
F11,3002 = 11,300  1,0002 > 12,000  1,0002 = 0.30
In either case, the probability is 0.30.
The expected value and variance of a uniform random variable X are computed as follows: a + b E3X4 = (5.20) 2
Var[X] =
1b  a2 2 (5.21) 12
A variation of the uniform distribution is one for which the random variable is restricted to integer values between a and b (also integers); this is called a discrete uniform distribution. An example of a discrete uniform distribution is the roll of a single die. Each of the numbers 1 through 6 has a 16 probability of occurrence.
Figure 5.15
Uniform Probability Density Function
1/1,000 $1,000
$1 , 1/1,000
50 $1 0 ,7 00 $2 ,0 00
Figure 5.17
P($1,500 6 X 6 $1,700)
$1 ,0 00
▶▶
00
00 0
1/1,000
$1 ,3
Figure 5.16
Probability that X 6 $1,300
$1 ,
▶▶
$2,000
$2 ,0 00
▶▶
Chapter 5 Probability Distributions and Data Modeling
227
Normal Distribution The normal distribution is a continuous distribution that is described by the familiar bellshaped curve and is perhaps the most important distribution used in statistics. The normal distribution is observed in many natural phenomena. Test scores such as the SAT, deviations from specifications of machined items, human height and weight, and many other measurements are often normally distributed. The normal distribution is characterized by two parameters: the mean, m, and the standard deviation, s. Thus, as m changes, the location of the distribution on the xaxis also changes, and as s is decreased or increased, the distribution becomes narrower or wider, respectively. Figure 5.18 shows some examples. The normal distribution has the following properties: 1. The distribution is symmetric, so its measure of skewness is zero. 2. The mean, median, and mode are all equal. Thus, half the area falls above the mean and half falls below it. 3. The range of X is unbounded, meaning that the tails of the distribution extend to negative and positive infinity. 4. The empirical rules apply exactly for the normal distribution; the area under the density function within {1 standard deviation is 68.3%, the area under the density function within {2 standard deviations is 95.4%, and the area under the density function within {3 standard deviations is 99.7%. Normal probabilities cannot be computed using a mathematical formula. Instead, we may use the Excel function NORM.DIST(x, mean, standard_deviation, cumulative). NORM.DIST(x, mean, standard_deviation, TRUE) calculates the cumulative probability F1x2 = P1X … x2 for a specified mean and standard deviation. (If cumulative is set to FALSE, the function simply calculates the value of the density function f(x), which has little practical application other than tabulating values of the density function. This was used to draw the distributions in Figure 5.18 using Excel.)
▶▶
Figure 5.18
Examples of Normal Distributions
228
Chapter 5 Probability Distributions and Data Modeling
EXAMPLE 5.34 Using the NORM.DIST Function to Compute Normal Probabilities Suppose that a company has determined that the distribution of customer demand (X) is normal with a mean of 750 units/month and a standard deviation of 100 units/ month. Figure 5.19 shows some cumulative probabilities calculated with the NORM.DIST function (see the Excel file Normal Probabilities). The company would like to know the following:
1. What is the probability that demand will be at most 900 units? 2. What is the probability that demand will exceed 700 units? 3. What is the probability that demand will be between 700 and 900 units? To answer the questions, first draw a picture. This helps to ensure that you know what area you are trying to calculate and how to use the formulas for working with a cumulative distribution correctly. ■■
Question 1. Figure 5.20(a) shows the probability that demand will be at most 900 units, or P(X … 900).
■■
This is simply the cumulative probability for x = 900, which can be calculated using the Excel function = NORM.DIST(900, 750, 100, TRUE) = 0.9332. Question 2. Figure 5.20(b) shows the probability that demand will exceed 700 units, P(X 7 700). Using the principles we have previously discussed, this can be found by subtracting P(X 6 700) from 1: P(X 7 700) = 1  P(X 6 700) = 1  F(700) = 1  0.3085 = 0.6915
This can be computed in Excel using the formula = 1  NORM.DIST(700, 750, 100, TRUE). ■■ Question 3. The probability that demand will be between 700 and 900, P(700 … X … 900), is illustrated in Figure 5.20(c). This is calculated by P(700 … X … 900) = P (X … 900)  P(X … 700) = F(900)  F (700) = 0.9332  0.3085 = 0.6247 In Excel, we would use the formula = NORM.DIST (900, 750, 100, TRUE)  NORM.DIST(700, 750, 100, TRUE).
The NORM.INV Function With the NORM.DIST function, we are given a value, x, of the random variable X and can find the cumulative probability to the left of x. Now let’s reverse the problem. Suppose that we know the cumulative probability but don’t know the value of x. How can we find it? We are often faced with such a question in many applications. The Excel function NORM. INV(probability, mean, standard_dev) can be used to do this. In this function, probability is the cumulative probability value corresponding to the value of x we seek, and “INV” stands for inverse.
▶▶
Figure 5.19
Normal Probability Calculations in Excel
▶▶
Chapter 5 Probability Distributions and Data Modeling
Figure 5.20
229
P(X … 900)
Computing Normal Probabilities
u = 100
u = 100
P(X > 700) x
750
x
900
700 750
(a)
(b) u = 100
Area = 1  0.10
P(700 … X … 900)
u = 100 Area = 0.10
700 750
x
900
750
(c)
?
x
(d)
EXAMPL E 5.35 Using the NORM.INV Function In the previous example, what level of demand would be exceeded at most 10% of the time? Here, we need to find the value of x so that P (X 7 x) = 0.10. This is illustrated in Figure 5.20(d). Because the area in the upper tail of the normal distribution is 0.10, the cumulative probability must be 1  0.10 = 0.90. From Figure 5.19,
we can see that the correct value must be somewhere between 850 and 900 because F(850) = 0.8413 and F(900) = 0.9332. We can find the exact value using the Excel function = NORM.INV(0.90, 750, 100) = 878.155. Therefore, a demand of approximately 878 will satisfy the criterion.
Standard Normal Distribution Figure 5.21 provides a sketch of a special case of the normal distribution called the standard normal distribution—the normal distribution with m = 0 and s = 1. This distribution is important in performing many probability calculations. A standard normal random variable is usually denoted by Z, and its density function by f 1z2. The scale along the zaxis represents the number of standard deviations from the mean of zero. The Excel function NORM.S.DIST(z) finds probabilities for the standard normal distribution.
EXAMPL E 5.36 Computing Probabilities with the Standard Normal Distribution We have previously noted that the empirical rules apply to any normal distribution. Let us find the areas under the standard normal distribution within 1, 2, and 3 standard deviations of the mean. These can be found by using the function NORM.S.DIST(z). Figure 5.22 shows a tabulation of the cumulative probabilities for z ranging from 3 to +3 and calculations of the areas within 1, 2, and 3 standard deviations of the mean. We apply formula (5.17) to find the difference between the cumulative
probabilities, F (b)  F (a). For example, the area within 1 standard deviation of the mean is found by calculating P(1 6 Z 6 1) = F(1)  F(1) = NORM.S.DIST(1) NORM.S.DIST(1) = 0.84134  0.15866 = 0.6827 (the difference due to decimal rounding). As the empirical rules stated, about 68% of the area falls within 1 standard deviation; 95% within 2 standard deviations; and more than 99% within 3 standard deviations of the mean.
230 ▶▶
Chapter 5 Probability Distributions and Data Modeling
Figure 5.21
Standard Normal Distribution
▶▶
Figure 5.22
Computing Standard Normal Probabilities
Using Standard Normal Distribution Tables Although it is quite easy to use Excel to compute normal probabilities, tables of the standard normal distribution are commonly found in textbooks and professional references when a computer is not available. Such a table is provided in Table A.1 of Appendix A at the end of this book. The table allows you to look up the cumulative probability for any value of z between 3.00 and +3.00. One of the advantages of the standard normal distribution is that we may compute probabilities for any normal random variable X having a mean m and standard deviation s by converting it to a standard normal random variable Z. We introduced the concept of standardized values (zscores) for sample data in Chapter 4. Here, we use a similar formula to convert a value x from an arbitrary normal distribution into an equivalent standard normal value, z:
z =
x  m (5.22) s
Chapter 5 Probability Distributions and Data Modeling
231
EXAMPL E 5.37 Computing Probabilities with Standard Normal Tables We will answer the first question posed in Example 5.34: What is the probability that demand will be at most x = 900 units if the distribution of customer demand (X) is normal with a mean of 750 units/month and a standard deviation of 100 units/month? Using formula (5.22), convert x to a standard normal value: 900 750 = 1.5 z = 100
Note that 900 is 150 units higher than the mean of 750; since the standard deviation is 100, this simply means that 900 is 1.5 standard deviations above the mean, which is the value of z. Using Table A.1 in Appendix A, we see that the cumulative probability for z = 1.5 is 0.9332, which is the same answer we found for Example 5.34.
Exponential Distribution The exponential distribution is a continuous distribution that models the time between randomly occurring events. Thus, it is often used in such applications as modeling the time between customer arrivals to a service system or the time to or between failures of machines, lightbulbs, hard drives, and other mechanical or electrical components. Similar to the Poisson distribution, the exponential distribution has one parameter, l. In fact, the exponential distribution is closely related to the Poisson distribution; if the number of events occurring during an interval of time has a Poisson distribution, then the time between events is exponentially distributed. For instance, if the number of arrivals at a bank is Poissondistributed, say with mean l = 12>hour, then the time between arrivals is exponential, with mean m = 1>12 hour, or 5 minutes. The exponential distribution has the density function
f 1x2 = lelx, for x Ú 0 (5.23)
and its cumulative distribution function is
F1x2 = 1  elx, for x Ú 0 (5.24)
Sometimes, the exponential distribution is expressed in terms of the mean m rather than the rate l. To do this, simply substitute 1>m for l in the preceding formulas. The expected value of the exponential distribution is 1>l and the variance is 11>l2 2. Figure 5.23 provides a sketch of the exponential distribution. The exponential distribution has the properties that it is bounded below by 0, it has its greatest density at 0, and the density declines as x increases. The Excel function EXPON.DIST(x, lambda, cumulative) can be used to compute exponential probabilities. As with other Excel probability distribution functions, cumulative is either TRUE or FALSE, with TRUE providing the cumulative distribution function. EXAMPL E 5.38 Using the Exponential Distribution Suppose that the mean time to failure of a critical component of an engine is m = 8,000 hours. Therefore, l = 1>m = 1>8,000 failures/hour. The probability that the component will fail before x hours is given by the cumulative distribution function F(x). Figure 5.24 shows a
portion of the cumulative distribution function, which may be found in the Excel file Exponential Probabilities. For example, the probability of failing before 5,000 hours is F (5,000) = 0.4647.
232 ▶▶
Chapter 5 Probability Distributions and Data Modeling
Figure 5.23
Example of an Exponential Distribution (l = 1)
▶▶
Figure 5.24
Computing Exponential Probabilities in Excel
Triangular Distribution The triangular distribution is defined by three parameters: the minimum, a; maximum, b; and most likely, c. Outcomes near the most likely value have a higher chance of occurring than those at the extremes. By varying the most likely value, the triangular distribution can be symmetric or skewed in either direction, as shown in Figure 5.25. Because the triangular distribution can assume different shapes, it is useful to model a wide variety of phenomena. For example, triangular distribution is often used when no data are available to characterize an uncertain variable and the distribution must be estimated judgmentally. The mean of the triangular distribution is calculated as
Mean = (a + c + b)/3 (5.25)
▶▶
Chapter 5 Probability Distributions and Data Modeling
Figure 5.25
233
f (x)
Examples of Triangular Distributions (symmetric)
a
c
b
x
f(x)
(positively skewed)
a
c
b
x
f(x)
(negatively skewed)
a
c
b
x
The variance is
Variance = (a2 + c2 + b2  a * c  a * b  c * b)/18 (5.26)
CHECK YOUR UNDERSTANDING 1. H ow does a continuous probability distribution differ from a discrete probability distribution? What properties must it have? 2. State how to compute probabilities over intervals for continuous distributions. 3. Explain the uniform distribution and how it might be used. 4. What properties does a normal distribution have? xplain how to use Excel functions NORM.DIST and NORM.INV to compute normal 5. E probabilities. 6. W hat is a standard normal distribution? How can you convert a random variable X having an arbitrary normal distribution into a standard normal random variable Z? 7. What typical situations do the exponential and triangular distributions model?
Data Modeling and Distribution Fitting In many applications of business analytics, we need to collect sample data of important variables such as customer demand, purchase behavior, machine failure times, and service activity times, to name just a few, to gain an understanding of the distributions of
234
Chapter 5 Probability Distributions and Data Modeling
these variables. Using the tools we have studied, we may construct frequency distributions and histograms and compute basic descriptive statistical measures to better understand the nature of the data. However, sample data are just that—samples. Using sample data may limit our ability to predict uncertain events that may occur because potential values outside the range of the sample data are not included. A better approach is to identify the underlying probability distribution from which sample data come by “fitting” a theoretical distribution to the data and verifying the goodness of fit statistically. To select an appropriate theoretical distribution that fits sample data, we might begin by examining a histogram of the data to look for the distinctive shapes of particular distributions. For example, normal data are symmetric, with a peak in the middle. Exponential data are very positively skewed, with no negative values. This approach is not, of course, always accurate or valid, and sometimes it can be difficult to apply, especially if sample sizes are small. However, it may narrow the search down to a few potential distributions. Summary statistics, such as the mean, median, standard deviation, and coefficient of variation, often provide information about the nature of the distribution. For instance, normally distributed data tend to have a fairly low coefficient of variation (however, this may not be true if the mean is small). For normally distributed data, we would also expect the median and mean to be approximately the same. For exponentially distributed data, however, the median will be less than the mean. Also, we would expect the mean to be about equal to the standard deviation, or, equivalently, the coefficient of variation would be close to 1. We could also look at the skewness index. For instance, normal data are not skewed. The following examples illustrate some of these ideas. The examination of histograms and summary statistics might provide some idea of the appropriate distribution; however, a better approach is to analytically fit the data to the best type of probability distribution.
EXAMPLE 5.39 Analyzing Airline Passenger Data An airline operates a daily route between two mediumsized cities using a 70seat regional jet. The flight is rarely booked to capacity but often accommodates business travelers who book at the last minute at a high price. Figure 5.26 shows the number of passengers for a sample of 25 flights (Excel file Airline Passengers). The histogram shows a relatively symmetric distribution. The mean, median, and mode are all similar, although there is some degree of positive
skewness. From our discussion in Chapter 4 about the variability of samples, it is important to recognize that although the histogram in Figure 5.26 does not look perfectly normal, this is a relatively small sample that can exhibit a lot of variability compared with the population from which it is drawn. Thus, based on these characteristics, it would not be unreasonable to assume a normal distribution for the purpose of developing a predictive or prescriptive analytics model.
EXAMPLE 5.40 Analyzing Airport Service Times Figure 5.27 shows a portion of the data and statistical analysis of 812 samples of service times at an airport’s ticketing counter (Excel file Airport Service Times). The data certainly do not appear to be normal or uniform. The histogram appears to look like an exponential distribution, and this might be a reasonable choice. However, there is a
difference between the mean and standard deviation, which we would not expect to see for an exponential distribution. Some other exotic distribution that we have not introduced might be more appropriate. Identifying the best fitting distribution can be done using sophisticated software.
▶▶
Chapter 5 Probability Distributions and Data Modeling
235
Figure 5.26
Data and Statistics for Passenger Demand
▶▶
Figure 5.27
Airport Service Times Statistics
Goodness of Fit: Testing for Normality of an Empirical Distribution The basis for fitting data to a probability distribution is a statistical procedure called goodness of fit. Goodness of fit attempts to draw a conclusion about the nature of the distribution. For instance, in Example 5.39 we suggested that it might be reasonable to assume that the distribution of passenger demand is normal. Goodness of fit would provide objective, analytical support for this assumption. We can do this using a procedure called the chisquare goodness of fit test.
236
Chapter 5 Probability Distributions and Data Modeling
ChiSquare Goodness of Fit Test The chisquare goodness of fit test can be used to determine whether sample data are representative of some probability distribution. We will explain the theory behind this test in Chapter 7; here, you simply need to understand the calculations. Example 5.41 illustrates its application for a normal distribution. The chisquare test can be applied to many distributions; the only modification to the procedure that must be made is to use the proper function to find the cumulative probabilities of the observed frequency distribution. For example, if we suspect that data follow an exponential distribution, then use =EXPON.DIST(upper limit, lambda, TRUE), where lambda is 1/sample mean. For a uniform distribution, the probabilities can be computed easily using simple geometry, like we did in Example 5.33.
EXAMPLE 5.41 Determining Normality for Airline Passenger Demand Data In Example 5.39, we analyzed passenger demand data in the Airline Passengers Excel file and concluded that it would not be unreasonable to assume that the data stem from a normal distribution. To test for this statistically, we do the following (this is shown in Figure 5.28 and in the Excel file Chi Square Goodness of Fit Test): 1. Start with the frequency distribution of the data; these are the observed frequencies. In Figure 5.28 we show this, as well as the descriptive statistics, in columns H through N. 2. Find the cumulative probability corresponding to the upper limit for each bin, assuming a normal distribution with the sample mean and sample standard deviation of the data. This is found using the function = NORM.DIST (upper limit, sample mean, sample standard deviation, TRUE) and is shown in column C of the spreadsheet. 3. Find the probability associated with each bin using formula (5.17). See question 3 in Example 5.34 for using the NORM.DIST function. This is computed in column D of the spreadsheet.
▶▶
Figure 5.28
Calculations for ChiSquare Test for Normality
4. Multiply the bin probabilities by the number of observations (in this case, 25). These values are the expected frequencies that you would see if the normal distribution were the correct data model. These are computed in column E of the spreadsheet. 5. Compute the difference between the observed and expected frequencies for each bin, square the result, and divide by the expected frequency, (O  E)2/E, for each bin (see column F of the spreadsheet). 6. Sum the chisquare calculations for all these bins (cell F17). This sum is called the chisquare statistic. 7. Compute the chisquare critical value using the function = CHISQ.INV.RT(0.05, number of bins  3). This is shown in cell F18 in the spreadsheet. 8. Compare the chisquare statistic with the critical value. If the chisquare statistic is less than or equal to the critical value, then the data can be reasonably assumed to come from a normal distribution having the sample mean and standard deviation. If not, then conclude that the normal distribution is not appropriate to model the data.
Chapter 5 Probability Distributions and Data Modeling
237
The number of degrees of freedom used in the CHISQ.INV.RT function depends on the type of distribution you are fitting. In general, degrees of freedom is the number of bins minus one minus the number of parameters of the distribution that are estimated from the data. For the normal distribution, we need to estimate two parameters — the mean and standard deviation; thus, degrees of freedom equals the number of bins minus 3 (see step 7 in Example 5.41). For the exponential distribution, we need only estimate one parameter, lambda, so the degrees of freedom are the number of bins minus 2.
ANALYTICS IN PRACTICE: The Value of Good Data Modeling in Advertising
3
Victor Correira/Shutterstock.com
Victor Correira/Shutterstock.com
To illustrate the importance of identifying the correct distribution in decision modeling, we discuss an example in advertising.3 The amount that companies spend on the creative component of advertising (that is, making better ads) is traditionally quite small relative to the overall media budget. One expert noted that the expenditure on creative development was about 5% of that spent on the media delivery campaign. Whatever money is spent on creative development is usually directed through a single advertising agency. However, one theory that has been proposed is that more should be spent on creative ad development, and the expenditures should be spread across a number of competitive advertising agencies. In research studies of this theory, the distribution of advertising effectiveness was assumed to be normal. In reality, data collected on the response to consumer product ads show that this distribution is actually quite skewed and, therefore, not normally distributed. Using the wrong assumption in any model or application can produce erroneous results. In this situation, the skewness actually provides an advantage for advertisers, making it more effective to obtain ideas from a variety of advertising agencies. A mathematical model (called Gross’s model) relates the relative contributions of creative and media dollars to total advertising effectiveness and is often used to identify the best number of draft ads to purchase. This model includes factors of ad development cost, total media spending budget, the distribution of effectiveness across ads (assumed to be normal), and the unreliability of identifying the most effective ad from a set of independently generated alternatives. Gross’s model concluded that large gains were possible if
multiple ads were obtained from independent sources, and the best ad is selected. Since the data observed on ad effectiveness was clearly skewed, other researchers examined ad effectiveness by studying standard industry data on ad recall without requiring the assumption of normally distributed effects. This analysis found that the best of a number of ads was more effective than any single ad. Further analysis revealed that the optimal number of ads to commission can vary significantly, depending on the shape of the distribution of effectiveness for a single ad. The researchers developed an alternative to Gross’s model. From their analyses, they found that as the number of draft ads was increased, the effectiveness of the best ad also increased. Both the optimal number of draft ads and the payoff from creating multiple independent drafts were higher when the correct distribution was used than the results reported in Gross’s original study.
Based on G. C. O’Connor, T. R. Willemain, and J. MacLachlan, “The Value of Competition Among Agencies in Developing Ad Campaigns: Revisiting Gross’s Model,” Journal of Advertising, 25, 1 (1996): 51–62.
238
Chapter 5 Probability Distributions and Data Modeling
CHECK YOUR UNDERSTANDING 1. Explain the concept of “fitting” a probability distribution to data. 2. What are some simple approaches to help you fit a distribution? 3. E xplain how to use the chisquare goodness of fit test to determine whether sample data can be adequately represented by a particular probability distribution. KEY TERMS Bernoulli distribution Binomial distribution Chisquare critical value Chisquare goodness of fit test Chisquare statistic Combination Complement Conditional probability Continuous random variable Cumulative distribution function Discrete random variable Discrete uniform distribution Empirical probability distribution Event Expected value Experiment Exponential distribution Goodness of fit Independent events Intersection
Joint probability Joint probability table Marginal probability Multiplication law of probability Mutually exclusive Normal distribution Outcome Permutation Poisson distribution Probability Probability density function Probability distribution Probability mass function Random variable Sample space Standard normal distribution Tree diagram Triangular distribution Uniform distribution Union
CHAPTER 5 TECHNOLOGY HELP Useful Excel Functions BINOM.DIST(number _s, trials, probability _s, cumulative) Computes probabilities for the binomial distribution.
CHISQ.INV.RT(probability, deg_freedom) Returns the value of chisquare that has a righttail area equal to probability for a specified degree of freedom. Used for the chisquare test
POISSON.DIST(x, mean, cumulative) Computes probabilities for the Poisson distribution. NORM.DIST(x, mean, standard_deviation, cumulative) Computes probabilities for the normal distribution. NORM.INV(probability, mean, standard_dev) Finds the value of x for a normal distribution having a specified cumulative probability. NORM.S.DIST(z) Finds probabilities for the standard normal distribution. EXPON.DIST(x, lambda, cumulative) Computes probabilities for the exponential distribution.
StatCrunch StatCrunch provides graphical calculations for all the distributions we studied in this chapter (binomial, discrete uniform, exponential, normal, Poisson, uniform, and custom, which is a userdefined discrete distribution) as well as many other advanced distributions. You can find video tutorials and stepbystep procedures with examples at https://www.statcrunch.com/5.0/example.php. We suggest that you first view the tutorials Getting started with StatCrunch and Working with StatCrunch sessions.
Chapter 5 Probability Distributions and Data Modeling
The following tutorials are listed under the heading Graphical Calculators: ■■ ■■
Continuous distributions Discrete distributions
Example: Using Graphical Calculators 1. Select the name of the desired distribution from the menu listing (for example, Binomial, Normal). 2. In the first line below the plot in the calculator window, specify the distribution parameters. As examples, with the normal distribution, specify the mean and standard deviation; with the binomial distribution, specify n and p.
239
3. In the second line below the plot, specify the direction of the desired probability. a. To compute a probability, enter a value to the right of the direction selector and leave the remaining field empty (for example, P(X 6 3) = ). b. To determine the point that will provide a specified probability, enter the probability to the right of the equal sign and leave the other field empty (for example, P(X 6 ) = 0.25). This option is available only for continuous distributions. 4. Click Compute! to fill in the empty fields and to update the graph of the distribution.
PROBLEMS AND EXERCISES b. Define the outcomes of the sample space for
Basic Concepts of Probability 1. In the past week, an icecream kiosk sold 15 cups of
chocolate flavor, 22 cups of vanilla, and 13 cups of strawberry. a. Estimate the probability of sale of each flavor for the following week. b. Define the definition of probability used to determine the answers in part a. 2. The students at Skövde University, Sweden, have
elected five candidates to represent them on the college management council. The committee summarized the age and gender of the members as follows: Member
Gender
Age
1
Male
18
2
Male
19
3
Female
22
4
Female
20
5
Male
23
This group of student representatives decide to elect a spokesperson by randomly drawing a name from a box. What is the probability that the spokesperson will be either a female or older than 21 years? 3. A glass jar contains 10 balls with labels of numbers 1
to 10. Two balls are randomly selected from the glass jar without replacement. a. Determine the number of combinations possible in the sample space. List all possible elements of the sample space.
which the sum of the selected balls is an odd number. c. What is the probability of obtaining all the outcomes (each sum of two balls) from part a? 4. A box contains marbles of four different colors: red,
green, blue, and yellow. Three marbles are randomly chosen from the box. a. List all possible outcomes in the sample space. b. What is the probability of each outcome? 5. The latest nationwide political poll in a particular
country indicates that the probability for the candidate to be a republican is 0.55, a communist is 0.30, and a supporter of the patriots of that country is 0.15. Assuming that these probabilities are accurate, within a randomly chosen group of 10 citizens: a. What is the probability that four are communists? b. What is the probability that none is a republican? 6. Refer to the ball scenario described in Problem 3. a. Let A be the event with the sum of the selected
two balls is an odd number. Find P(A) and P(Ac). b. What is the probability that the sum of the two balls will not be more than 5? c. What is the probability that the sum of the two balls will be between 5 and 7?
240
Chapter 5 Probability Distributions and Data Modeling
7. Answer the following by referring to the scenario
described in Problem 4. a. Let A be the event “without blue marble.” Find P(A). b. Let B be the event “first marble selected is green.” Find P(B). c. Let C be the event “last marble selected is red.” Find P(C). d. Are events A and B mutually exclusive? Find P(A or B). e. Are events A and C mutually exclusive? Find P(A or C)c. 8. Roulette is played at a table similar to the one in
Figure 5.29. A wheel with the numbers 1 through 36 (evenly distributed with the colors red and black) and two green numbers, 0 and 00, rotate in a shallow bowl with a curved wall. A small ball is spun on the inside of the wall and drops into a pocket corresponding to one of the numbers. Players may make 11 different types of bets by placing chips on different areas of the table. These include bets on a single number, two adjacent numbers, a row of three numbers, a block of four numbers, two adjacent rows of six numbers, and the five number combinations of 0, 00, 1, 2, and 3; bets on the numbers 1–18 or 19–36; the first, second, or third group of 12 numbers; a column of 12 numbers; even or odd; and red or black. Payoffs differ by bet. For instance, a singlenumber bet pays 35 to 1 if it wins; a threenumber bet pays 11 to 1; a column bet pays 2 to 1; and a color bet pays even money. Define the following events: C1 = column 1 number, C2 = column 2 number, C3 = column 3 number, O = odd number, E = even number, G = green number, F12 = first 12 numbers, S12 = second 12 numbers, and T12 = third 12 numbers. ▶▶
Figure 5.29
Layout of a Typical Roulette Table
a. Find the probability of each of these events. b. Find P(G or O), P(O or F12), P(C1 or C3), P(E and
F12), P(E or F12), P(S12 and T12), and P(O or C2). 9. Students in the new MBA class at a state university
have the following specialization profile: Finance—83 Marketing—36 Operations and Supply Chain Management—72 Information Systems—59 Find the probability that a student is either a finance or a marketing major. Are the events finance specialization and marketing specialization mutually exclusive? If so, what assumption must be made? 10. Sue and Dave always spend their evenings together,
either cycling or playing computer games. On an evening with good weather, the probability that Sue prefers cycling is 0.7 and for Dave it is 0.4. Find the probability that on a particular fine evening, a. both Sue and Dave prefer cycling. b. both Sue and Dave prefer playing computer games. c. either Sue or Dave (not both) prefers playing computer games. 11. A grocery shop is taking stock of its inventory. The
shopkeeper found that out of 120 tins with damaged labels, there are 50 tins of beans, 20 tins of sardines, 35 tins of processed meat, and the rest are tins of soup. a. Find the probability that the tins with damaged labels are all soup tins. b. What is the probability that the tins with damaged labels are not sardine tins? c. If two tins are selected, what is the probability that the tins with damaged labels are the tins of beans and the meat tins?
Chapter 5 Probability Distributions and Data Modeling
12. Row 26 of the Excel file Census Education Data
14. A Canadian business school summarized the gender
gives the number of employed persons in the civilian labor force having a specific educational level. a. Find the probability that an employed person has attained each of the educational levels listed in the data. b. Suppose that A is the event “An employed person has some type of college degree” and B is the event “An employed person has at least some college.” Find the probabilities of these events. Are they mutually exclusive? Why or why not? c. Find the probability P(A or B). Explain what this means.
and residency of its incoming class as follows: Residency Gender Canada United States Europe Asia Other Male 125 18 17 50 8 Female 103 8 10 92 4 a. Construct a joint probability table. b. Calculate the marginal probabilities. c. What is the probability that a female student is
from outside Canada or the United States? 15. In Example 4.13, we developed the following cross
tabulation of sales transaction data:
13. A survey of shopping habits found the percentage
of respondents that use technology for shopping as shown in Figure 5.30. For example, 17.39% only use online coupons; 21.74% use online coupons and check prices online before shopping, and so on.
Region
a. What is the probability that a shopper will check
West Total
b. c. d. e.
f. g.
▶▶
prices online before shopping? What is the probability that a shopper will use a smart phone to save money? What is the probability that a shopper will use online coupons? What is the probability that a shopper will not use any of these technologies? What is the probability that a shopper will check prices online and use online coupons but not use a smart phone? If a shopper checks prices online, what is the probability that he or she will use a smart phone? What is the probability that a shopper will check prices online but not use online coupons or a smart phone?
Book
DVD
Total
East
56
42
98
North
43
42
85
South
62
37
99
100
90
190
261
211
472
a. Find the marginal probabilities that a sale origi
nated in each of the four regions and the marginal probability of each type of sale (book or DVD). b. Find the conditional probabilities of selling a book given that the customer resides in each region. 16. Use the Civilian Labor Force data in the Excel file
Census Education Data to find the following: a. P(unemployed and advanced degree) b. P(unemployed ∙ advanced degree) c. P(not a high school grad ∙ unemployed) d. Are the events “unemployed” and “at least a high school graduate” independent?
Figure 5.30 4.35%
Check Prices Online Before Shopping
21.74%
17.39%
241
17.39%
4.35%
Use Online Coupons
4.35%
4.35% Use a Smart Phone to Save Money
242
Chapter 5 Probability Distributions and Data Modeling
17. Using the data in the Excel file Consumer Trans
port Survey, develop a crosstabulation for Gender and Vehicle Driven; then convert this table into probabilities. a. What is the probability that a respondent is female? b. What is the probability that a respondent drives an SUV? c. What is the probability that a respondent is male and drives a minivan? d. What is the probability that a female respondent drives either a truck or an SUV? e. If it is known that an individual drives a car, what is the probability that the individual is female? f. If it is known that an individual is male, what is the probability that he drives an SUV? g. Determine whether the random variable “gender”
and the event “vehicle driven” are statistically independent. What would this mean for advertisers?
of successful and unsuccessful products given known test market results, that is, P(successful product  favorable test market), P(successful product  unfavorable test market), P(unsuccessful p roduct  favorable test market), and P(unsuccessful product  unfavorable test market). Discrete Probability Distributions 22. An investor estimates that there is a 1 in 10 chance
that a stock purchase will lose 20% of its value, a 2 in 10 chance that it will break even, a 3 in 10 chance that it will gain 15%, and a 4 in 10 chance that it will gain 30%. What is the expected return as a percentage based on these estimates? 23. The number of children per family was determined
and summarized in the following table. Number of Children
Number of families
1
15
2
31
3
24
4
7
18. A home pregnancy test is not always accurate. Sup
pose the probability is 0.015 that the test indicates that a woman is pregnant when she actually is not, and the probability is 0.025 that the test indicates that a woman is not pregnant when she really is. Assume that the probability that a woman who takes the test is actually pregnant is 0.7. What is the probability that a woman is pregnant if the test yields a notpregnant result? 19. Based on the scenario in Problem 3, what is the prob
ability of first selecting ball 1 and then selecting ball 3? Will it be different if the first ball is replaced in the glass jar? Explain the difference between these two scenarios. 20. In the roulette example described in Problem 8, what
is the probability that the outcome will be green twice in a row? What is the probability that the outcome will be black twice in a row?
Find the expected number, variance, and standard deviation of the number of children per family. 24. Padini Concept, a local retail store in Malaysia,
recorded the number of shoes purchased by customers (1 pairs, 2 pairs and 3 pairs) based on the number of clothes purchased (1, 2, and 3) for a random sample of 250 customers. The following data was recorded: Number of Customers
Pairs of Shoes
62
1
80
3
Clothes
21. A consumer products company found that 44% of
110
successful products also received favorable results from test market research, whereas 11% had unfavorable results but nevertheless were successful. That is, P(successful product and favorable test market) = 0.44 and P(successful product and unfavorable test market) = 0.11. They also found that 32% of unsuccessful products had unfavorable research results, whereas 13% of them had favorable research results, that is P(unsuccessful product and unfavorable test market) = 0.32 and P(unsuccessful product and favorable test market) = 0.13. Find the probabilities
21
2
1
18
3
1
31
1
2
24
3
2
32
2
3
2
a. What is the probability of customers purchasing 2
pairs of shoes? b. What is the probability of customers purchasing 3
items of clothing?
Chapter 5 Probability Distributions and Data Modeling
c. Construct a cumulative probability distribution
for the purchase of shoes. What is the probability of the number of shoes purchased being less than 3? What is the probability of number of pairs of shoes purchased being between 2 and 3? d. Find the expected value and standard deviation for the number of pairs of shoes purchased. 25. From a bag full of colored balls (red, blue, green, and
243
will be occupied? Use the binomial distribution to find the exact value. 31. A transport company is inspecting 80 trucks it owns.
The inspection determines that the probability of trucks failing the inspection is 0.25. Find the probability that not more than 10 trucks failed the inspection. What is the probability that 75 or more trucks passed the inspection?
orange), a few are picked out and replaced. This is done 1,000 times and the number of times each colored ball is picked out is, blue balls = 300, red balls = 200, green balls = 450, and orange balls = 50. a. What is the probability of picking a green ball? b. What is the probability of picking a blue ball?
32. A life insurance consultant sells three life insur
c. If there are 100 balls in the bag, how many of
c. one policy, assuming that there are five working
them are likely to be green? d. If there are 10,000 balls in the bag, how many of them are likely to be orange? 26. A major application of analytics in marketing is
determining customer retention. Suppose that the probability of a longdistance carrier’s customer leaving for another carrier from one month to the next is 0.16. What distribution models the retention of an individual customer? What is the expected value and standard deviation? 27. The Excel file Call Center Data shows that in a
sample of 70 individuals, 27 had prior call center experience. If we assume that the probability that any potential hire will also have experience with a probability of 27/70, what is the probability that among ten potential hires, more than half of them will have experience? Define the parameter(s) for this distribution based on the data.
ance policies on an average per week for which she receives a fee. Use Poisson’s distribution to calculate the probability that in a given week she will sell a. some policies. b. two or more policies but less than five policies. days a week. 33. The number and frequency of hurricanes over a
certain ocean, annually from 1945 through 2020, is shown here. For instance, this means that no hurricanes occurred during 6 of these years, only one hurricane occurred in 16 of these years, and so on. Number
Frequency
0
6
1
16
2
20
3
12
4
4
5
5
6
4
7
4
8
2
28. If a cell phone company conducted a telemarketing
11
1
campaign to generate new clients and the probability of successfully gaining a new customer was 0.07, what is the probability that contacting 50 potential customers would result in at least 5 new customers?
12
1
29. During one 12week period, a particular mutual fund
outperformed the S&P 500 index 37 out of 60 days. Find the probability that it would perform as well or better again. 30. A popular resort hotel has 300 rooms and is usually
fully booked. About 7% of the time a reservation is canceled before the 6:00 p.m. deadline with no penalty. What is the probability that at least 285 rooms
a. Find the probabilities of 0–12 hurricanes each
season using these data. b. Find the mean number of hurricanes. c. Assuming a Poisson distribution and using the mean number of hurricanes per season from part b, compute the probabilities of experiencing 0–12 hurricanes in a season. Compare these to your answer to part a. How accurately does a Poisson distribution model this phenomenon? Construct a chart to visualize these results.
244
Chapter 5 Probability Distributions and Data Modeling
Continuous Probability Distributions 34. Verify that the function corresponding to the fol
lowing figure is a valid probability density function. Then find the following probabilities: a. P(x 6 8) b. P(x 7 7) c. P(6 6 x 6 10) d. P(8 6 x 6 11)
38. The distribution of SAT scores in math for an incom
ing class of business students has a mean of 610 and standard deviation of 20. Assume that the scores are normally distributed. a. Find the probability that an individual’s SAT score is less than 600. b. Find the probability that an individual’s SAT score is between 590 and 620. c. Find the probability that an individual’s SAT score is greater than 650. d. What scores will the top 5% of students have? e. Find the standardized values for students scoring 540, 600, 650, and 700 on the test. Explain what these mean. 39. A popular soft drink is sold in 2liter (2,000milliliter)
35. The time required to play a game of Battleship™ is
uniformly distributed between 20 and 60 minutes. a. Find the expected value and variance of the time to complete the game. b. What is the probability of finishing within 30 minutes? c. What is the probability that the game would take longer than 40 minutes? 36. A student is drawing a rectangle, which has a perim
eter of 20 cm. He wants the length, x cm, of one side of this rectangle to be uniformly distributed between 1 cm and 7 cm. Find the probability that the length of the longer side of the rectangle is more than 6 cm long. 37. In determining bike mileage ratings, it was found that
the mpg (X) for a certain model is normally distributed, with a mean of 34 mpg and a standard deviation of 1.9 mpg. Find the following: a. P(X 6 33) b. P(31 6 X 6 38) c. P(X 7 36) d. P(X 7 33) e. The mileage rating that the upper 6% of bikes achieve.
bottles. Because of variation in the filling process, bottles have a mean of 2,000 milliliters and a standard deviation of 18, normally distributed. a. If the process fills the bottle by more than 30 milliliters, the overflow will cause a machine malfunction. What is the probability of this occurring? b. What is the probability of underfilling the bottles by at least 10 milliliters? 40. A fisherman found that the average length of cherry
hybrid salmon is about 20 inches with a standard deviation of 2 inches. By assuming the cherry salmon’s length is normally distributed, find the probability that its length is a. more than 25 inches. b. less than 17 inches. c. within 5% of the cherry salmon’s mean length. 41. According to historical data, a European musician
finds that customers who download music from a popular Web service spend approximately €24 per month, with a standard deviation of €4. Find the probability that a customer will spend at least €22 per month. How much (or more) do the top 9% of customers spend? 42. According to WHO, around 1.35 million people
die each year due to road accidents. Suppose that on a specific highway, an average of four accidents occur within every hour. Find the probability that the duration between accidents is less than or equal to 30 minutes.
Chapter 5 Probability Distributions and Data Modeling
245
43. The actual delivery time from a pizza delivery com
45. Apply the chisquare goodness of fit test to the data
pany is exponentially distributed with a mean of 28 minutes. a. What is the probability that the delivery time will exceed 31 minutes? b. What proportion of deliveries will be completed within 25 minutes?
in the Airport Service Times Excel file to determine if an exponential distribution models the data. Use bins of width 100. Note that for an exponential distribution, the number of degrees of freedom for the CHISQ.INV.RT function should be the number of bins minus 2 as discussed in the chapter. 46. Compute the daily change of the closing price for the
Data Modeling and Distribution Fitting 44. Apply the chisquare goodness of fit test to the data
in the Airport Service Times Excel file to determine if a normal distribution models the data. Use bins of width 100. Note that for a normal distribution, the number of degrees of freedom for the CHISQ.INV. RT function should be the number of bins minus 3 as discussed in the chapter.
data in the Excel file S&P 500. Compute descriptive statistics, a frequency distribution, and histogram for the closing prices (using a bin width of 25). What probability distribution would you propose as a good fit for the data? Verify your choice using the chisquare goodness of fit test.
CASE: PERFORMANCE LAWN EQUIPMENT PLE collects a variety of data from special studies, many of which are related to the quality of its products. The company collects data about functional test performance of its mowers after assembly; results from the past 30 days are given in the worksheet Mower Test in the Performance Lawn Equipment Database. In addition, many inprocess measurements are taken to ensure that manufacturing processes remain in control and can produce according to design specifications. The worksheet Blade Weight shows 350 measurements of blade weights taken from the manufacturing process that produces mower blades during the most recent shift. Elizabeth Burke has asked you to study these data from an analytics perspective. Drawing upon your experience, you have developed a number of questions. 1. For the mower test data, what distribution might be appropriate to model the failure of an individual mower? 2. What fraction of mowers fails the functional performance test using all the mower test data? 3. What is the probability of having x failures in the next 100 mowers tested, for x from 0 to 20?
4. What is the average blade weight and how much variability is occurring in the measurements of blade weights? 5. Assuming that the data are normal, what is the probability that blade weights from this process will exceed 5.20? 6. What is the probability that blade weights will be less than 4.80? 7. What is the actual percent of blade weights that exceed 5.20 or are less than 4.80 from the data in the worksheet? 8. Is the process that makes the blades stable over time? That is, are there any apparent changes in the pattern of the blade weights? 9. Could any of the blade weights be considered outliers, which might indicate a problem with the manufacturing process or materials? 10. Is the assumption that blade weights are normally distributed justified? Summarize all your findings to these questions in a wellwritten report.
This page intentionally left blank
CHAPTER
Sampling and Estimation
6
Robert Brown Stock/Shutterstock
LEAR NING OBJE CTIVE S ■■ ■■
■■ ■■
■■
■■
After studying this chapter, you will be able to:
Describe the elements of a sampling plan. Explain the difference between subjective and probabilistic sampling. State two types of subjective sampling. Explain how to conduct simple random sampling and use Excel to find a simple random sample from an Excel database. Explain systematic, stratified, and cluster sampling and sampling from a continuous process. Explain the importance of unbiased estimators.
■■
■■
■■ ■■ ■■
■■
Describe the difference between sampling error and nonsampling error. Explain how the average, standard deviation, and distribution of means of samples change as the sample size increases. Define the sampling distribution of the mean. Calculate the standard error of the mean. Explain the practical importance of the central limit theorem. Use the standard error in probability calculations.
248 ■■
■■ ■■
■■
Chapter 6 Sampling and Estimation
Explain how an interval estimate differs from a point estimate. Define and give examples of confidence intervals. Calculate confidence intervals for population means and proportions using the formulas in the chapter and the appropriate Excel functions. Explain how confidence intervals change as the level of confidence increases or decreases.
■■
■■
■■
■■
Describe the difference between the tdistribution and the normal distribution. Use and visualize confidence intervals to draw conclusions about population parameters. Compute a prediction interval and explain how it differs from a confidence interval. Compute sample sizes needed to ensure a confidence interval for means and proportions with a specified margin of error.
We discussed the difference between population and samples in Chapter 4. Sampling is the foundation of statistical analysis. We use sample data in business analytics applications for many purposes. For example, we might wish to estimate the mean, variance, or proportion of a very large or unknown population; provide values for inputs in decision models; understand customer satisfaction; reach a conclusion as to which of several sales strategies is most effective; or understand if a change in a process resulted in an improvement. In this chapter, we discuss sampling methods, how they are used to estimate population parameters, and how we can assess the error inherent in sampling. In your career, you will discover many opportunities to apply statistical sampling and estimation. For example, marketing analysts routinely use surveys to understand customer demographics, brand and product attribute preferences, and customer satisfaction. Accounting professionals use sampling techniques to audit invoices, accounts receivable, and freight bills. Operations managers use sampling to verify the quality of incoming materials and production output. Human resource managers use surveys and sampling techniques to evaluate employee satisfaction. Companies in the entertainment industry routinely use sampling to rate TV shows, movies, and video games. So, every time you are asked to complete a survey, you are seeing sampling and estimation at work!
Statistical Sampling The first step in sampling is to design an effective sampling plan that will yield representative samples of the populations under study. A sampling plan is a description of the approach that is used to obtain samples from a population prior to any data collection activity. A sampling plan states ■■ ■■ ■■
the objectives of the sampling activity, the target population, the population frame (the list from which the sample is selected),
Chapter 6 Sampling and Estimation
■■ ■■ ■■
EXAMPL E 6.1
249
the method of sampling, the operational procedures for collecting the data, and the statistical tools that will be used to analyze the data.
A Sampling Plan for a Market Research Study
Suppose that a company wants to understand how golfers might respond to a membership program that provides discounts at golf courses in the golfers’ locality as well as across the country. The objective of a sampling study might be to estimate the proportion of golfers who would likely subscribe to this program. The target population might be all golfers over 25 years old. However, identifying all golfers in America might be impossible. A practical population frame might be a list of golfers who have
purchased equipment from national golf or sporting goods companies through which the discount card will be sold. The operational procedures for collecting the data might be an email link to a survey site or directmail questionnaire. The data might be stored in an Excel database; statistical tools such as PivotTables and simple descriptive statistics would be used to segment the respondents into different demographic groups and estimate their likelihood of responding positively.
Sampling Methods Many types of sampling methods exist. Sampling methods can be subjective or probabilistic. Subjective methods include judgment sampling, in which expert judgment is used to select the sample (survey the “best” customers), and convenience sampling, in which samples are selected based on the ease with which the data can be collected (survey all customers who happen to visit this month). Probabilistic sampling involves selecting the items in the sample using some random procedure. Probabilistic sampling is necessary to draw valid statistical conclusions. The most common probabilistic sampling approach is simple random sampling. Simple random sampling involves selecting items from a population so that every subset of a given size has an equal chance of being selected. If the population data are stored in a database, simple random samples can generally be easily obtained.
EXAMPL E 6.2
Simple Random Sampling with Excel
Suppose that we wish to sample from the Excel database Sales Transactions. Excel provides a tool to generate a random set of values from a given population size. Click on Data Analysis in the Analysis group of the Data tab and select Sampling. This brings up the dialog shown in Figure 6.1. In the Input Range box, we specify the data range from which the sample will be taken. This tool requires that the data sampled be numeric, so in this example we sample from the first column of the data set, which corresponds to the customer ID number. There are two options for sampling: 1. Sampling can be Periodic, and we will be prompted for the Period, which is the interval between sample
observations from the beginning of the data set. For instance, if a period of 5 is used, observations 5, 10, 15, and so on will be selected as samples. 2 . Sampling can also be Random, and we will be prompted for the Number of Samples. Excel will then randomly select this number of samples from the specified data set. However, this tool generates random samples with replacement, so we must be careful to check for duplicate observations in the sample created. Figure 6.2 shows 20 random samples generated by the tool. We sorted them in ascending order to make it easier to identify duplicates. As you can see, two of the customers were duplicated by the tool.
250
Chapter 6 Sampling and Estimation
▶▶ Figure
6.1
Excel Sampling Tool Dialog
▶▶ Figure
6.2
Samples Generated Using the Excel Sampling Tool
Other methods of sampling include the following: ■■
■■
Systematic (Periodic) Sampling. Systematic, or periodic, sampling is a sampling plan (one of the options in the Excel Sampling tool) that selects every nth item from the population. For example, to sample 250 names from a list of 400,000, the first name could be selected at random from the first 1,600, and then every 1,600th name could be selected. This approach can be used for telephone sampling when supported by an automatic dialer that is programmed to dial numbers in a systematic manner. However, systematic sampling is not the same as simple random sampling because for any sample, every possible sample of a given size in the population does not have an equal chance of being selected. In some situations, this approach can induce significant bias if the population has some underlying pattern. For instance, sampling orders received every seven days may not yield a representative sample if customers tend to send orders on certain days every week. Stratified Sampling. Stratified sampling applies to populations that are divided into natural subsets (called strata) and allocates the appropriate proportion of
Chapter 6 Sampling and Estimation
■■
■■
251
samples to each stratum. For example, a large city may be divided into political districts called wards. Each ward has a different number of citizens. A stratified sample would choose a sample of individuals in each ward proportionate to its size. This approach ensures that each stratum is weighted by its size relative to the population and can provide better results than simple random sampling if the items in each stratum are not homogeneous. However, issues of cost or significance of certain strata might make a disproportionate sample more useful. For example, the ethnic or racial mix of each ward might be significantly different, making it difficult for a stratified sample to obtain the desired information. Cluster Sampling. Cluster sampling is based on dividing a population into subgroups (clusters), sampling a set of clusters, and (usually) conducting a complete census within the clusters sampled. For instance, a company might segment its customers into small geographical regions. A cluster sample would consist of a random sample of the geographical regions, and all customers within these regions would be surveyed (which might be easier because regional lists might be easier to produce and mail). Sampling from a Continuous Process. Selecting a sample from a continuous manufacturing process can be accomplished in two main ways. First, select a time at random; then select the next n items produced after that time. Second, select n times at random; then select the next item produced after each of these times. The first approach generally ensures that the observations will come from a homogeneous population; however, the second approach might include items from different populations if the characteristics of the process should change over time, so caution should be used. CHECK YOUR UNDERSTANDING
1. State the major elements of a sampling plan. 2. What is simple random sampling? 3. Describe and give an example of systematic, stratified, cluster, and continuous process sampling methods.
ANALYTICS IN PRACTICE: Using Sampling Techniques to Improve
Distribution1
U.S. breweries rely on a threetier distribution system to deliver product to retail outlets, such as supermarkets and convenience stores, and onpremise accounts, such as bars and restaurants. The three tiers are the manufacturer, wholesaler (distributor), and retailer. A distribution network must be as efficient and cost
effective as possible to deliver to the market a fresh product that is damage free and is delivered at the right place at the right time. To understand distributor performance related to overall effectiveness, MillerCoors brewery defined seven attributes of proper distribution and collected data from 500 of
(continued ) 1
Based on Tony Gojanovic and Ernie Jimenez, “Brewed Awakening: Beer Maker Uses Statistical Methods to Improve How Its Products Are Distributed,” Quality Progress (April 2010).
Chapter 6 Sampling and Estimation
its distributors. A field quality specialist (FQS) audits distributors within an assigned region of the country and collects data on these attributes. The FQS uses a handheld device to scan the universal product code on each package to identify the product type and amount. When audits are complete, data are summarized and uploaded from the handheld device into a master database. This distributor auditing uses stratified random sampling with proportional allocation of samples based on the distributor’s market share. In addition to providing a more representative sample and better logistical control of sampling, stratified random sampling enhances statistical precision when data are aggregated by market area served by the distributor. This enhanced precision is a consequence of smaller and typically homogeneous market regions, which are able to provide realistic estimates of variability, especially when compared to another market region that is markedly different. Randomization of retail accounts is achieved through a specially designed program based on the GPS location of the distributor and serviced retail accounts. The sampling strategy ultimately addresses a specific distributor’s performance related to outofcode product, damaged product, and outofrotation product at the retail level. All in all, more than 6,000 of the brewery’s national retail
Stephen Finn/Shutterstock.com
252
accounts are audited during a sampling year. Data collected by the FQSs during the year are used to develop a performance ranking of distributors and identify opportunities for improvement.
Estimating Population Parameters Sample data provide the basis for many useful analyses to support decision making. Estimation involves assessing the value of an unknown population parameter—such as a population mean, population proportion, or population variance—using sample data. Estimators are the measures used to estimate population parameters; for example, we use the sample mean x to estimate a population mean m. The sample variance s2 estimates a population variance s2, and the sample proportion p estimates a population proportion p. A point estimate is a single number derived from sample data that is used to estimate the value of a population parameter.
Unbiased Estimators It seems quite intuitive that the sample mean should provide a good point estimate for the population mean. However, it may not be clear why the formula for the sample variance that we introduced in Chapter 4 has a denominator of n  1, particularly because it is different from the formula for the population variance (see formulas (4.7) and (4.8) in Chapter 4). In these formulas, the population variance is computed by 2 a (xi  m) n
s2 =
i=1
N
Chapter 6 Sampling and Estimation
253
whereas the sample variance is computed by the formula 2 a (xi  x) n
2
s =
i=1
n  1
Why is this so? Statisticians develop many types of estimators, and from a theoretical as well as a practical perspective, it is important that they “truly estimate” the population parameters they are supposed to estimate. Suppose that we perform an experiment in which we repeatedly sampled from a population and computed a point estimate for a population parameter. Each individual point estimate will vary from the population parameter; however, we would hope that the longterm average (expected value) of all possible point estimates would equal the population parameter. If the expected value of an estimator equals the population parameter it is intended to estimate, the estimator is said to be unbiased. If this is not true, the estimator is called biased and will not provide correct results. Fortunately, all the estimators we have introduced are unbiased and, therefore, are meaningful for making decisions involving the population parameter. In particular, statisticians have shown that the denominator n  1 used in computing s2 is necessary to provide an unbiased estimator of s2. If we simply divided by the number of observations, the estimator would tend to underestimate the true variance.
Errors in Point Estimation One of the drawbacks of using point estimates is that they do not provide any indication of the magnitude of the potential error in the estimate. A major metropolitan newspaper reported that, based on a Bureau of Labor Statistics survey, college professors were the highestpaid workers in the region, with an average salary of $150,004. Actual averages for two local universities were less than $70,000. What happened? As reported in a followup story, the sample size was very small and included a large number of highly paid medical school faculty; as a result, there was a significant error in the point estimate that was used. When we sample, the estimators we use—such as a sample mean, sample proportion, or sample variance—are actually random variables that are characterized by some distribution. By knowing what this distribution is, we can use probability theory to quantify the uncertainty associated with the estimator. To understand this, we first need to discuss sampling error and sampling distributions. In Chapter 4, we observed that different samples from the same population have different characteristics—for example, variations in the mean, standard deviation, frequency distribution, and so on. Sampling (statistical) error occurs because samples are only a subset of the total population. Sampling error is inherent in any sampling process, and although it can be minimized, it cannot be totally avoided. Another type of error, called nonsampling error, occurs when the sample does not represent the target population adequately. This is generally a result of poor sample design, such as using a convenience sample when a simple random sample would have been more appropriate, or choosing the wrong population frame. It may also result from inadequate data reliability, which we discussed in Chapter 1. To draw good conclusions from samples, analysts need to eliminate nonsampling error and understand the nature of sampling error. Sampling error depends on the size of the sample relative to the population. Thus, determining the number of samples to take is essentially a statistical issue that is based on the accuracy of the estimates needed to draw a useful conclusion. We discuss this later in this chapter. However, from a practical standpoint, one must also consider the cost of sampling and sometimes make a tradeoff between cost and the information that is obtained.
254
Chapter 6 Sampling and Estimation
Understanding Sampling Error Suppose that we estimate the mean of a population using the sample mean. How can we determine how accurate we are? In other words, can we make an informed statement about how far the sample mean might be from the true population mean? We could gain some insight into this question by performing a sampling experiment as Example 6.3 illustrates. If we apply the empirical rules to these results, we can estimate the sampling error associated with one of the sample sizes we have chosen (see Example 6.4). EXAMPLE 6.3
A Sampling Experiment
Let us choose a population that is uniformly distributed between a = 0 and b = 10. Formulas (5.20) and (5.21) state that the expected value (population mean) is (0 + 10)>2 = 5, and the variance is (10  0)2 >12 = 8.333. We use the Excel Random Number Generation tool described in Chapter 5 to generate 25 samples, each of size ten, from this population. Figure 6.3 shows a portion of a spreadsheet for this experiment, along with a histogram of the data (on the left side) that shows that the 250 observations are approximately uniformly distributed. (This is available in the Excel file Sampling Experiment.) In row 12, we compute the mean of each sample. These statistics vary a lot from the population values because of sampling error. The histogram on the right shows the distribution of the 25 sample means, which vary from less than 4 to more than 6. Now let’s compute the average and standard deviation of the sample means in row 12 (cells
▲▲ Figure
6.3
Portion of Spreadsheet for Sampling Experiment
AB12 and AB13). Note that the average of all the sample means is quite close to the true population mean of 5.0. Now let us repeat this experiment for larger sample sizes. Table 6.1 shows some results. Notice that as the sample size gets larger, the averages of the 25 sample means are all still close to the expected value of 5; however, the standard deviation of the 25 sample means becomes smaller for increasing sample sizes, meaning that the means of samples are clustered closer together around the true expected value. Figure 6.4 shows comparative histograms of the sample means for each of these cases. These illustrate the conclusions we just made and, perhaps even more surprisingly, the distribution of the sample means appears to assume the shape of a normal distribution for larger sample sizes. In our experiment, we used only 25 sample means. If we had used a muchlarger number, the distributions would have been more well defined.
Chapter 6 Sampling and Estimation
▶▶ Table
6.1
Results from Sampling Experiment
▲▲ Figure
Sample Size
Average of 25 Sample Means
Standard Deviation of 25 Sample Means
10
5.0108
0.816673
25
5.0779
0.451351
100
4.9137
0.301941
500
4.9754
0.078993
255
6.4
Histograms of Sample Means for Increasing Sample Sizes
EXAMPL E 6.4 Estimating Sampling Error Using the Empirical Rules Using the standard deviations in Table 6.1 in Example 6.3 and the empirical rule for three standard deviations around the mean, we could state, for example, that using a sample size of ten, the distribution of sample means should fall within three standard deviations of the population mean, 5, or approximately from 5.0  3(0.816673) = 2.55 to 5.0 + 3(0.816673) = 7.45. Thus, there is considerable
error in estimating the mean using a sample of only ten. For a sample of size 25, we would expect the sample means to fall between 5.0  (0.451351) = 3.65 and 5.0 + 3(0.451351) = 6.35. Note that as the sample size increased, the error decreased. For sample sizes of 100 and 500, the intervals are [4.09, 5.91] and [4.76, 5.24].
CHECK YOUR UNDERSTANDING 1. What is an unbiased estimator, and why is it important in statistics? 2. Explain the difference between sampling and nonsampling error. 3. What happens to the distribution of the sample mean as the sample size increases?
256
Chapter 6 Sampling and Estimation
Sampling Distributions We can quantify the sampling error in estimating the mean for any unknown population. To do this, we need to characterize the sampling distribution of the mean.
Sampling Distribution of the Mean The means of all possible samples of a fixed size n from some population will form a distribution that we call the sampling distribution of the mean. The histograms in Figure 6.4 are approximations to the sampling distributions of the mean based on 25 s amples. Statisticians have shown two key results about the sampling distribution of the mean. First, the standard deviation of the sampling distribution of the mean, called the standard error of the mean, is computed as
Standard Error of the Mean = s> 1n (6.1)
where s is the standard deviation of the population from which the individual observations are drawn and n is the sample size. From this formula, we see that as n increases, the standard error decreases, just as our experiment demonstrated. This suggests that the estimates of the mean that we obtain from larger sample sizes provide greater accuracy in estimating the true population mean. In other words, larger sample sizes have less sampling error.
EXAMPLE 6.5 Computing the Standard Error of the Mean For our experiment, we know that the variance of the population is 8.33 (because the values were uniformly distributed). Therefore, the standard deviation of the population is s = 2.89. We may compute the standard error of the mean for each of the sample sizes in our experiment using formula (6.1). For example, with n = 10, we have Standard Error of the Mean = s> 2n = 2.89> 210 = 0.914
For the remaining data in Table 6.1 we have the following:
Sample Size, n
Standard Error of the Mean
10
0.914
25
0.5778
100
0.289
500
0.129
The standard deviations shown in Table 6.1 are simply estimates of the standard error of the mean based on using only 25 samples. If we compare these estimates with the theoretical values in the previous example, we see that they are close but not exactly the same. This is because the true standard error is based on all possible sample means in the sampling distribution, whereas we used only 25. If you repeat the experiment with a larger number of samples, the observed values of the standard error would be closer to these theoretical values. In practice, we will never know the true population standard deviation and generally take only a limited sample of n observations. However, we may estimate the standard error of the mean using the sample data by simply dividing the sample standard deviation by the square root of n. The second result that statisticians have shown is called the central limit theorem, one of the most important practical results in statistics that makes systematic inference possible. The central limit theorem states that if the sample size is large enough, the sampling distribution of the mean is approximately normally distributed, regardless of the distribution of the population, and that the mean of the sampling distribution will be the same
Chapter 6 Sampling and Estimation
257
as that of the population. This is exactly what we observed in our experiment. The distribution of the population was uniform, yet the sampling distribution of the mean converges to the shape of a normal distribution as the sample size increases. The central limit theorem also states that if the population is normally distributed, then the sampling distribution of the mean will also be normal for any sample size. The central limit theorem allows us to use the theory we learned about calculating probabilities for normal distributions to draw conclusions about sample means.
Applying the Sampling Distribution of the Mean The key to applying sampling distribution of the mean correctly is to understand whether the probability that you wish to compute relates to an individual observation or to the mean of a sample. If it relates to the mean of a sample, then you must use the sampling distribution of the mean, whose standard deviation is the standard error, s> 1n. EXAMPL E 6.6 Using the Standard Error in Probability Calculations Suppose that the size of individual customer orders (in dollars), X, from a major discount book publisher Web site is normally distributed with a mean of $36 and standard deviation of $8. The probability that the next individual who places an order at the Web site will make a purchase of more than $40 can be found by calculating 1  NORM.DIST(40, 36, 8, TRUE) = 1  0.6915 = 0.3085
Now suppose that a sample of 16 customers is chosen. What is the probability that the mean purchase for these 16 customers will exceed $40? To find this, we must realize that we must use the sampling distribution of the mean to carry out the appropriate calculations. The sampling distribution
of the mean will have a mean of $36 but a standard error of $8> 216 = $2. Then the probability that the mean purchase exceeds $40 for a sample size of n = 16 is 1  NORM.DIST(40, 36, 2, TRUE) = 1  0.9772 = 0.0228
Although about 30% of individuals will make purchases exceeding $40, the chance that 16 customers will collectively average more than $40 is much smaller. It would be very unlikely for all 16 customers to make highvolume purchases, because some individual purchases would as likely be less than $36 as more, making the variability of the mean purchase amount for the sample of 16 much smaller than that for individuals.
CHECK YOUR UNDERSTANDING 1. What is the sampling distribution of the mean? 2. Define the standard error of the mean, and explain how to compute it. 3. What does the central limit theorem state, and of what value is it? 4. Explain when to use the standard error in probability calculations versus using the standard deviation of the population.
Interval Estimates An interval estimate provides a range for a population characteristic based on a sample. Intervals are quite useful in statistics because they provide more information than a point estimate. Intervals specify a range of plausible values for the characteristic of interest and a way of assessing “how plausible” they are. In general, a 100(1  a)% probability interval is any interval [A, B] such that the probability of falling between A and B is 1  a. Probability intervals are often centered on the mean or median. For instance, in a normal
258
Chapter 6 Sampling and Estimation
distribution, the mean plus or minus 1 standard deviation describes an approximate 68% probability interval around the mean. As another example, the 5th and 95th percentiles in a data set constitute a 90% probability interval. EXAMPLE 6.7 Interval Estimates in the News We see interval estimates in the news all the time when trying to estimate the mean or proportion of a population. Interval estimates are often constructed by taking a point estimate and adding and subtracting a margin of error that is based on the sample size. For example, a Gallup poll might report that 56% of voters support a certain candidate with a margin of error of { 3%. We would conclude that the true percentage of
voters that support the candidate is most likely between 53% and 59%. Therefore, we would have a lot of confidence in predicting that the candidate would win a forthcoming election. If, however, the poll showed a 52% level of support with a margin of error of { 4%, we might not be as confident in predicting a win because the true percentage of supportive voters is likely to be somewhere between 48% and 56%.
The question you might be asking at this point is how to calculate the error associated with a point estimate. In national surveys and political polls, such margins of error are usually stated, but they are never properly explained. To understand them, we need to introduce the concept of confidence intervals.
Confidence Intervals Confidence interval estimates provide a way of assessing the accuracy of a point estimate. A confidence interval is a range of values between which the value of the population parameter is believed to be, along with a probability that the interval correctly estimates the true (unknown) population parameter. This probability is called the level of confidence, denoted by 1  a, where a is a number between 0 and 1. The level of confidence is usually expressed as a percent; common values are 90%, 95%, or 99%. (Note that if the level of confidence is 90%, then a = 0.1.) The margin of error depends on the level of confidence and the sample size. For example, suppose that the margin of error for some sample size and a level of confidence of 95% is calculated to be 2.0. One sample might yield a point estimate of 10. Then, a 95% confidence interval would be [8, 12]. However, this interval may or may not include the true population mean. If we take a different sample, we will most likely have a different point estimate, say, 10.4, which, given the same margin of error, would yield the interval estimate [8.4, 12.4]. Again, this may or may not include the true population mean. If we chose 100 different samples, leading to 100 different interval estimates, we would expect that 95% of them—the level of confidence—would contain the true population mean. We would say we are “95% confident” that the interval we obtain from sample data contains the true population mean. The higher the confidence level, the more assurance we have that the interval contains the true population parameter. As the confidence level increases, the confidence interval becomes wider to provide higher levels of assurance. You can view a as the risk of incorrectly concluding that the confidence interval contains the true mean. When national surveys or political polls report an interval estimate, they are actually confidence intervals. However, the level of confidence is generally not stated because the average person would probably not understand the concept or terminology. While not stated, you can probably assume that the level of confidence is 95%, as this is the most common value used in practice (however, the Bureau of Labor Statistics tends to use 90% quite often). Many different types of confidence intervals may be developed. The formulas used depend on the population parameter we are trying to estimate and possibly other characteristics or assumptions about the population. We illustrate a few types of confidence intervals.
Chapter 6 Sampling and Estimation
259
Confidence Interval for the Mean with Known Population Standard Deviation The simplest type of confidence interval is for the mean of a population where the standard deviation is assumed to be known. You should realize, however, that in nearly all practical sampling applications, the population standard deviation will not be known. However, in some applications, such as measurements of parts from an automated machine, a process might have a very stable variance that has been established over a long history, and it can reasonably be assumed that the standard deviation is known. A 100(1  a)% confidence interval for the population mean m based on a sample of size n with a sample mean x and a known population standard deviation s is given by
x { za/2(s> 1n) (6.2)
Note that this formula is simply the sample mean (point estimate) plus or minus a margin of error. The margin of error is a number za>2 multiplied by the standard error of the sampling distribution of the mean, s> 1n. The value za>2 represents the value of a standard normal random variable that has an upper tail probability of a>2 or, equivalently, a cumulative probability of 1  a>2. It may be found from the standard normal table (see Table A.1 in Appendix A at the end of the book) or may be computed in Excel using the value of the function NORM.S.INV(1  a>2). For example, if a = 0.05 (for a 95% confidence interval), then =NORM.S.INV(0.975) = 1.96; if a = 0.10 (for a 90% confidence interval), then =NORM.S.INV(0.95) = 1.645, and so on. Although formula (6.2) can easily be implemented in a spreadsheet, the Excel function CONFIDENCE.NORM(alpha, standard_deviation, size) can be used to compute the margin of error term, za>2 s> 1n; thus, the confidence interval is the sample mean { CONFIDENCE.NORM(alpha, standard_deviation, size).
EXAMPL E 6.8
Computing a Confidence Interval with a Known Standard Deviation
In a production process for filling bottles of liquid detergent, historical data have shown that the variance in the volume is constant; however, clogs in the filling machine often affect the average volume. The historical standard deviation is 15 milliliters. In filling 800milliliter bottles, a sample of 25 found an average volume of 796 milliliters. Using formula (6.2), a 95% confidence interval for the population mean is
▶▶ Figure
6.5
Confidence Interval for Mean Liquid Detergent Filling Volume
x { za/2 (s> 1n)
= 796 { 1.96(15> 225) = 796 { 5.88, or [790.12, 801.88]
The worksheet Population Mean Sigma Known in the Excel workbook Confidence Intervals computes this interval using the CONFIDENCE.NORM function to compute the margin of error in cell B9, as shown in Figure 6.5.
260
Chapter 6 Sampling and Estimation
As the level of confidence, 1  a, decreases, za>2 decreases, and the confidence interval becomes narrower. For example, a 90% confidence interval will be narrower than a 95% confidence interval. Similarly, a 99% confidence interval will be wider than a 95% confidence interval. Essentially, you must trade off a higher level of accuracy with the risk that the confidence interval does not contain the true mean. Smaller risk will result in a wider confidence interval. However, you can also see that as the sample size increases, the standard error decreases, making the confidence interval narrower and providing a more accurate interval estimate for the same level of risk. So if you wish to reduce the risk, you should consider increasing the sample size.
The tDistribution In most practical applications, the standard deviation of the population is unknown, and we need to calculate the confidence interval differently. Before we can discuss how to compute this type of confidence interval, we need to introduce a new probability distribution called the tdistribution. The tdistribution is actually a family of probability distributions with a shape similar to the standard normal distribution. Different tdistributions are distinguished by an additional parameter, degrees of freedom (df). The tdistribution has a larger variance than the standard normal, thus making confidence intervals wider than those obtained from the standard normal distribution, in essence, correcting for the uncertainty about the true standard deviation, which is not known. As the number of degrees of freedom increases, the tdistribution converges to the standard normal distribution (Figure 6.6). When sample sizes get to be as large as 120, the distributions are virtually identical; even for sample sizes as low as 30 to 35, it becomes difficult to distinguish between the two. Thus, for large sample sizes, many people use zvalues as in formula (6.2) to establish confidence intervals even when the population standard deviation is unknown but estimated from the sample data. We must point out, however, that for any sample size, the true sampling distribution of the mean is the tdistribution; so when in doubt, use the tdistribution as we will illustrate in the next section. The concept of degrees of freedom can be puzzling. It can best be explained by examining the formula for the sample variance: 2 a (xi  x) n
s2 =
i=1
n  1
Note that to compute s2, we first need to compute the sample mean, x. If we know the value of the mean, then we need know only n  1 distinct observations; the nth is completely determined. (For instance, if the mean of three numbers is 4 and you know that two of the numbers are 2 and 4, you can easily determine that the third number must be 6.) The ▶▶ Figure
6.6
Comparison of the tDistribution to the Standard Normal Distribution
Chapter 6 Sampling and Estimation
261
number of sample values that are free to vary defines the number of degrees of freedom; in general, df equals the number of sample values minus the number of estimated parameters. Because the sample variance uses one estimated parameter, the mean, the tdistribution used in confidence interval calculations has n  1 degrees of freedom. Because the tdistribution explicitly accounts for the effect of the sample size in estimating the population variance, it is the proper one to use for any sample size. However, for large samples, the difference between t and zvalues is very small, as we noted earlier.
Confidence Interval for the Mean with Unknown Population Standard Deviation The formula for a 100(1  a)% confidence interval for the mean m when the population standard deviation is unknown is x { ta>2,n  1(s> 1n) (6.3)
where ta>2,n  1 is the value from the tdistribution with n  1 degrees of freedom, giving an uppertail probability of a>2. We may find tvalues in Table A.2 in Appendix A at the end of the book or by using the Excel function T.INV(1  a>2, n  1) or the function T.INV.2T(a, n  1). The Excel function CONFIDENCE.T(alpha, standard_deviation, size) can be used to compute the margin of error term, ta>2,n  1(s> 1n); thus, the confidence interval is the sample mean {CONFIDENCE.T(alpha, standard_deviation, size). EXAMPL E 6.9 Computing a Confidence Interval with Unknown Standard Deviation In the Excel file Credit Approval Decisions, a large bank has sample data used in making credit approval decisions (see Figure 6.7). Suppose that we want to find a 95% confidence interval for the mean revolving balance for the population of applicants who own a home. First, sort the data by homeowner and compute the mean and standard deviation of the revolving balance for the sample of homeowners. This results in x = $12,630.37 and s = $5,393.38. The sample
size is n = 27, so the standard error s> 2n = $1,037.96. The tdistribution has 26 degrees of freedom; therefore, t.025,26 = 2.056. Using formula (6.3), the confidence interval is $12,630.37 { 2.056($1,037.96) or [$10,497, $14,764]. The worksheet Population Mean Sigma Unknown in the Excel workbook Confidence Intervals computes this interval using the CONFIDENCE.T function to compute the margin of error in cell B10, as shown in Figure 6.8.
Confidence Interval for a Proportion For categorical variables such as gender (male or female), education (high school, college, postgraduate), and so on, we are usually interested in the proportion of observations in a sample that has a certain characteristic. An unbiased estimator of a population
▲▲ Figure
6.7
Portion of Excel File Credit Approval Decisions
262
Chapter 6 Sampling and Estimation
▶▶ Figure
6.8
Confidence Interval for Mean Revolving Balance of Homeowners
proportion p (this is not the number pi = 3.14159 . . . ) is the statistic pn = x>n (the sample proportion), where x is the number in the sample having the desired characteristic and n is the sample size. A 100(1  a)% confidence interval for the proportion is pn { za>2
pn(1  pn) (6.4) n B
Notice that as with the mean, the confidence interval is the point estimate plus or minus some margin of error. In this case, 3pn(1  pn)>n is the standard error for the sampling distribution of the proportion. Excel does not have a function for computing the margin of error, but it can easily be implemented on a spreadsheet. The value of za>2 is found using the Excel function NORM.S.INV(1  a>2). EXAMPLE 6.10
Computing a Confidence Interval for a Proportion
The last column in the Excel file Insurance Survey (see Figure 6.9) describes whether a sample of employees would be willing to pay a lower premium for a higher deductible for their health insurance. Suppose we are interested in the proportion of individuals who answered yes. We may easily confirm that 6 out of the 24 employees, or 25%, answered yes. Thus, a point estimate for the proportion answering yes is pn = 0.25. Using formula (6.4), we find that a 95% confidence interval for the proportion of employees answering yes is
▲▲ Figure
6.9
Portion of Excel File Insurance Survey
0.25(0.75) = 0.25 { 0.173, or 30.077, 0.4234 B 24 The worksheet Proportion in the Excel workbook Confidence Intervals computes this interval, as shown in Figure 6.10. Notice that this is a fairly wide confidence interval, suggesting that we have quite a bit of uncertainty as to the true value of the population proportion. This is because of the relatively small sample size. 0.25 { 1.96
Chapter 6 Sampling and Estimation
▶▶ Figure
263
6.10
Confidence Interval for the Proportion
Additional Types of Confidence Intervals Confidence intervals may be calculated for other population parameters such as a variance or standard deviation and also for differences in the means or proportions of two populations. The concepts are similar to the types of confidence intervals we have discussed, but many of the formulas are rather complex and more difficult to implement on a spreadsheet. Some advanced software packages and spreadsheet addins provide additional support. Therefore, we do not discuss them in this book, but we do suggest that you consult other books and statistical references should you need to use them, now that you understand the basic concepts underlying them. CHECK YOUR UNDERSTANDING 1. What is an interval estimate? 2. Explain the difference between a probability interval and a confidence interval. 3. What does the level of confidence specify? 4. Explain the differences in the formulas used to compute confidence intervals for the mean with known and unknown standard deviations. 5. How does the tdistribution compare with the normal distribution?
Using Confidence Intervals for Decision Making Confidence intervals can be used in many ways to support business decisions. EXAMPL E 6.11 Drawing a Conclusion About a Population
Mean Using a Confidence Interval In packaging a commodity product such as laundry detergent, the manufacturer must ensure that the packages contain the stated amount to meet government regulations. In Example 6.8, we saw an example where the required volume is 800 milliliters, yet the sample average was only 796 milliliters. Does this indicate a serious problem? Not necessarily. The 95% confidence interval for the mean we computed in Figure 6.5 was [790.12, 801.88]. Although the sample mean is less than 800, the sample does not provide sufficient evidence to draw the conclusion that the population mean is less than 800 because 800 is contained within
the confidence interval. In fact, it is just as plausible that the population mean is 801. We cannot tell definitively because of the sampling error. However, suppose that the sample average is 792. Using the Excel worksheet Population Mean Sigma Known in the workbook Confidence Intervals, we find that the confidence interval for the mean would be [786.12, 797.88]. In this case, we would conclude that it is highly unlikely that the population mean is 800 milliliters because the confidence interval falls completely below 800; the manufacturer should check and adjust the equipment to meet the standard.
264
Chapter 6 Sampling and Estimation
The next example shows how to interpret a confidence interval for a proportion. EXAMPLE 6.12
Using a Confidence Interval to Predict Election Returns
Suppose that an exit poll of 1,300 voters found that 692 voted for a particular candidate in a twoperson race. This represents a proportion of 53.23% of the sample. Could we conclude that the candidate will likely win the election? A 95% confidence interval for the proportion is [0.505, 0.559]. This suggests that the population proportion of voters who favor this candidate is highly likely to exceed 50%, so it is safe to predict the winner. On the other hand,
suppose that only 670 of the 1,300 voters voted for the candidate, a sample proportion of 0.515. The confidence interval for the population proportion is [0.488, 0.543]. Even though the sample proportion is larger than 50%, the sampling error is large, and the confidence interval suggests that it is reasonably likely that the true population proportion could be less than 50%, so it would not be wise to predict the winner based on this information.
Data Visualization for Confidence Interval Comparison We may use an Excel stock chart to visualize confidence intervals easily (the stock chart requires that you have at least three variables). This is particularly useful for comparing confidence intervals for different groups. EXAMPLE 6.13
Creating a Stock Chart for Confidence Intervals
Suppose that we constructed confidence intervals for monthly credit card debt for a sample of banking customers in four age groups. The data are shown below. Age Group
Upper CI
Lower CI
Mean
25–34 35–44 45–54 55–64
$2,103 $1,617 $1,114 $1,931
$711 $872 $468 $309
$1,407.14 $1,244.44 $ 791.30 $1,120.00
The columns should correspond to the variable name, upper CI limit, lower CI limit, and mean, in that order. Next, highlight the range of this table and insert an Excel HighLowClose Stock Chart. Rightclick one of the confidence intervals in the chart and choose Format HighLow Lines . . . . In the dropdown menu for HighLow Line Options, choose one of the series (Upper CI, Lower CI, or Mean) to format the markers. In the Format Data Series pane, click the paint icon and then Marker, making sure to expand the Marker Options menu. Choose the type of marker you wish and ▶▶ Figure
6.11
Visualization of Confidence Intervals Using an Excel Stock Chart
increase the width of the markers to make them more visible. We chose a “ + ” for the upper CI limit, a dash for the Lower CI limit, and a diamond for the mean. This results in the chart shown in Figure 6.11. We may use confidence intervals and their visualizations to determine whether the means of two populations are significantly different from each other. The rule is simple: If the confidence intervals overlap, then you cannot conclude statistically that the means differ. However, if the confidence intervals do not overlap, then you can conclude that there is a statistical difference in the means. In Example 6.13, for instance, suppose we wish to determine whether the mean of the 35–44 age group is significantly different from that of the 45–54 age group. In Figure 6.11, we see that the confidence intervals overlap, so we cannot conclude that the means are significantly different. In the next chapter, we will discuss these concepts in a more formalized manner using a statistical approach called hypothesis testing.
Chapter 6 Sampling and Estimation
265
CHECK YOUR UNDERSTANDING 1. Explain how to use confidence intervals to draw conclusions about the population mean or proportion. 2. How can confidence interval visualization be used to draw conclusions about a population?
Prediction Intervals Another type of interval used in estimation is a prediction interval. A prediction interval is one that provides a range for predicting the value of a new observation from the same population. This is different from a confidence interval, which provides an interval estimate of a population parameter, such as the mean or proportion. A confidence interval is associated with the sampling distribution of a statistic, but a prediction interval is associated with the distribution of the random variable itself. When the population standard deviation is unknown, a 100(1  a)% prediction interval for a new observation is x { ta>2,n  1 ¢ s
A
1 +
1 ≤ (6.5) n
Note that this interval is wider than the confidence interval in formula (6.3) because of the additional value of 1 under the square root. This is because, in addition to estimating the population mean, we must also account for the variability of the new observation around the mean. One important thing to realize also is that in formula (6.3) for a confidence interval, as n gets large, the error term tends to zero so the confidence interval converges on the mean. However, in the prediction interval formula (6.5), as n gets large, the error term converges to ta>2, n  1(s), which is simply a 100(1  a)% probability interval. Because we are trying to predict a new observation from the population, there will always be uncertainty.
EXAMPL E 6.14
Computing a Prediction Interval
In estimating the revolving balance in the Excel file Credit Approval Decisions in Example 6.9, we may use formula (6.5) to compute a 95% prediction interval for the revolving balance of a new homeowner as $12,630.37 { 2.056($5,393.38) 3$1,338.10, $23,922.644
A
1 +
Note that compared with Example 6.9, the size of the prediction interval is considerably wider than that of the confidence interval.
1 , or 27
CHECK YOUR UNDERSTANDING 1. How does a prediction interval differ from a confidence interval? 2. Why are prediction intervals wider than confidence intervals?
266
Chapter 6 Sampling and Estimation
Confidence Intervals and Sample Size An important question in sampling is the size of the sample to take. Note that in all the formulas for confidence intervals, the sample size plays a critical role in determining the width of the confidence interval. As the sample size increases, the width of the confidence interval decreases, providing a more accurate estimate of the true population parameter. In many applications, we would like to control the margin of error in a confidence interval. For example, in reporting voter preferences, we might wish to ensure that the margin of error is {2%. Fortunately, it is relatively easy to determine the appropriate sample size needed to estimate the population parameter within a specified level of precision. The formulas for determining sample sizes to achieve a given margin of error are based on the confidence interval halfwidths. For example, consider the confidence interval for the mean with a known population standard deviation we introduced in formula (6.2): x { za>2 ¢
s 2n
≤
Suppose we want the width of the confidence interval on either side of the mean (that is, the margin of error) to be at most E. In other words, E Ú za>2 ¢ Solving for n, we find:
n Ú (za>2)2
s 2n s2 E2
≤
(6.6)
Of course, we generally do not know the population standard deviation prior to finding the sample size. A commonsense approach would be to take an initial sample to estimate the population standard deviation using the sample standard deviation s and determine the required sample size, collecting additional data if needed. If the halfwidth of the resulting confidence interval is within the required margin of error, then we clearly have achieved our goal. If not, we can use the new sample standard deviation s to determine a new sample size and collect additional data as needed. Note that if s changes significantly, we still might not have achieved the desired precision and might have to repeat the process. Usually, however, this will be unnecessary. In a similar fashion, we can compute the sample size required to achieve a desired confidence interval halfwidth for a proportion by solving the following equation (based on formula (6.4) using the population proportion p in the margin of error term) for n: E Ú za>2 2p(1  p)>n This yields
n Ú (za>2)2
p(1  p) E2
(6.7)
Chapter 6 Sampling and Estimation
267
In practice, the value of p will not be known. You could use the sample proportion from a preliminary sample as an estimate of p to plan the sample size, but this might require several iterations and additional samples to find the sample size that yields the required precision. When no information is available, the most conservative estimate is to set p = 0.5. This maximizes the quantity p(1  p) in the formula, resulting in the sample size that will guarantee the required precision no matter what the true proportion is. EXAMPL E 6.15
Sample Size Determination for the Mean
In the liquid detergent example (Example 6.8), the confidence interval we computed in Figure 6.5 was [790.12, 801.88]. The width of the confidence interval is { 5.88 milliliters, which represents the sampling error. Suppose the manufacturer would like the sampling error to be at most 3 milliliters. Using formula (6.6), we may compute the required sample size as follows: n Ú 1za>2 2 2 = 11.962
EXAMPL E 6.16
Rounding up, we find that 97 samples would be needed. To verify this, Figure 6.12 shows that if a sample of 97 is used along with the same sample mean and standard deviation, the confidence interval does indeed have a sampling error less than 3 milliliters.
(s2)
E2 (152) 2 32
= 96.04
Sample Size Determination for a Proportion
For the voting example we discussed, suppose that we wish to determine the number of voters to poll to ensure a sampling error of at most { 2%. As we stated, when no information is available, the most conservative approach is to use 0.5 for the estimate of the true proportion. Using formula (6.7) with p = 0.5, the number of voters to poll to obtain a 95%
confidence interval on the proportion of voters that choose a particular candidate with a precision of { 0.02 or less is n Ú 1za/2 2 2
p(1  p)
= 11.962 2
E2 (0.5) (1  0.5) 0.022
= 2,401
CHECK YOUR UNDERSTANDING 1. Explain why it is important to determine the proper sample size to use in a survey. 2. How do you find the sample size for a proportion if no good estimate of the population proportion is available? ▶▶ Figure
6.12
Confidence Interval for the Mean Using Sample Size = 97
268
Chapter 6 Sampling and Estimation
KEY TERMS Central limit theorem Cluster sampling Confidence interval Convenience sampling Degrees of freedom (df) Estimation Estimators Interval estimate Judgment sampling Level of confidence Nonsampling error Point estimate
Population frame Prediction interval Probability interval Sample proportion Sampling distribution of the mean Sampling plan Sampling (statistical) error Simple random sampling Standard error of the mean Stratified sampling Systematic (or periodic) sampling tDistribution
CHAPTER 6 TECHNOLOGY HELP Useful Excel Functions NORM.S.INV(1  a/2) Finds the value of za>2, a standard normal random variable that has an upper tail probability of a>2 or, equivalently, a cumulative probability of 1  a>2. Used for confidence intervals for the mean with a known population standard deviation. CONFIDENCE.NORM(alpha, standard_deviation, size) Computes the margin of error for a confidence interval for the mean with a known population standard deviation. T.INV(1  a>2, n  1) or T.INV.2T(a, n  1) Finds the value from the tdistribution with n  1 degrees of freedom, giving an uppertail probability of a>2. Used for confidence intervals for the mean with an unknown population standard deviation. CONFIDENCE.T(alpha, standard_deviation, size) Computes the margin of error for a confidence interval for the mean with an unknown population standard deviation. Excel Templates Confidence Intervals (Examples 6.8–6.10): Open the Excel file Confidence Intervals. Select the worksheet tab corresponding to the type of confidence interval to calculate and enter the appropriate data. StatCrunch StatCrunch provides calculations of confidence intervals. You can find video tutorials and stepbystep procedures with examples at https://www.statcrunch.com/5.0/example. php. We suggest that you first view the tutorials Getting started with StatCrunch and Working with StatCrunch sessions. The following tutorials, listed under Confidence
Intervals For on this Web page, explain how to calculate confidence intervals in StatCrunch: ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■
A mean with raw data A mean with summary data A proportion with raw data A proportion with summary data The difference between two means with raw data The difference between two means with summary data The difference between means with paired data The difference between two proportions with raw data The difference between two proportions with summary data
Example: Calculating a Confidence Interval for the Mean with Raw Data 1. Choose the Stat 7 T Stats 7 One Sample 7 With Data menu option. 2. Select the column containing the data. 3. Under Perform, choose Confidence interval for m. 4. Change the confidence level if needed and click Compute!. Example: Calculating a Confidence Interval for the Difference in Proportions with Summary Data 1. Choose the Stat 7 Proportion Stats 7 Two Sample 7 With Summary menu option. 2. Under Sample 1, enter the # of successes and the # of observations. Under Sample 2, enter the # of successes and the # of observations. 3. Under Perform, choose Confidence interval for p1  p2. 4. Change the confidence level if needed and click Compute!.
Chapter 6 Sampling and Estimation
269
PROBLEMS AND EXERCISES Statistical Sampling 1. Your college or university wishes to obtain reliable
information about student perceptions of administrative communication. Describe how to design a sampling plan for this situation based on your knowledge of the structure and organization of your college or university. How would you implement simple random sampling, stratified sampling, and cluster sampling for this study? What would be the pros and cons of using each of these methods? 2. The Excel file Diabetes Check contains the records of
patients with diabetes who have received treatment at a local hospital. Number each row in the dataset and help the hospital to select a simple random sample of 20 unique results by using the Excel Sampling tool. 3. Describe how to apply stratified sampling to sample
from the Credit Risk Data file based on the different types of loans. Implement your process in Excel to choose a random sample consisting of 10% of the records for each type of loan. 4. The Excel file High Blood Pressure records the
results of 30 individuals who were tested for hypertension. Use the Excel Sampling tool to obtain a random sample of 5 individuals and compute point estimates for the mean and standard deviation for the systolic blood pressure. Are there any differences between the parameter and statistic values? Why? Estimating Population Parameters 5. Repeat the sampling experiment in Example 6.3 for
sample sizes 50, 100, 250, and 500. Compare your results to the example and use the empirical rules to analyze the sampling error. For each sample, also find the standard error of the mean using formula (6.1). 6. Using the Excel file Diabetes Check, draw five ran
dom samples by repeating the sampling experiment for sample sizes 10, 15, and 20. For each sample sizes, find the stabilized glucose point estimates for the mean and standard deviation. Estimate the sampling error by using the empirical rules.
the test, what would be the standard error of the mean? 8. In one region in your country, suppose that the stan
dard deviation of monthly expenditure of the residents is about $7,000. In a random sample of 380 residents, what is the standard error of the mean monthly expenditure? 9. Suppose that the mean score for the mathematics test
cited in Problem 7 is 607. What is the probability that a random sample of 324 students will have a mean score of more than 619? Less than 588? 10. Suppose that the mean monthly expenditure of the
region identified in Problem 8 is $15,500. What is the probability that the random sample of 380 residents will have a mean monthly expenditure of more than $15,000? Less than $14,500? 11. In determining automobile mileage ratings, it was
found that the mpg in the city for a certain model is normally distributed, with a mean of 41 mpg and a standard deviation of 1.8 mpg. Suppose that the car manufacturer samples five cars from its assembly line and tests them for mileage ratings. a. What is the distribution of the mean mpg for the sample? b. What is the probability that the mean mpg of the sample will be greater than 42 mpg? c. What is the probability that the mean mpg of the sample will be less than 40.5 mpg? 12. Each can of a certain brand of sweet corn contains an
average of 375 corn kernels with a standard deviation of 8 in each can. Yesterday, Winnie went to a local store and purchased four cans of corn. a. Find the probability that the average number of corn kernels in the cans purchased by Winnie will exceed 382. b. What is the maximum average number of corn kernels in the cans purchased will not be more than 10%? Interval Estimates
Sampling Distributions 7. Based on extensive data from a national high school
educational testing program, the standard deviation of national test scores for mathematics was found to be 118 points. If a sample of 324 students are given
13. A random sample of 100 teenagers was surveyed,
and the mean number of songs that they had downloaded from the iTunes store in the past month was 9.4 with the results considered accurate being within 1.4 (18 times out of 20).
270
Chapter 6 Sampling and Estimation
a. What percent of confidence level is the result? b. What is the margin of error? c. What is the confidence interval? Explain. 14. A sample of 33 airline passengers found that the
average checkin time is 2.167 minutes. Based on longterm data, the population standard deviation is known to be 0.48 minutes. Find a 95% confidence interval for the mean checkin time. Use the appropriate formula and verify your result using the Confidence Intervals workbook. 15. A survey of 34 freshmen found that they get an aver
age of 6.26 hours of sleep each night. A 90% confidence interval had a margin of error of 0.509. a. What are the lower and upper limits of the confidence interval? b. What was the standard deviation, assuming that
the population standard deviation is known? 16. A sample of 20 international students attending an
urban U.S. university found that the average amount budgeted for expenses per month was $1,612.50 with a standard deviation of $1,179.64. Find a 95% confidence interval for the mean monthly expense budget of the population of international students. Use the appropriate formula and verify your result using the Confidence Intervals workbook. 17. A sample of 40 individuals at a shopping mall found
that the mean number of visits to a restaurant per week was 2.88 with a standard deviation of 1.59. Find a 99% confidence interval for the mean number of restaurant visits. Use the appropriate formula and verify your result using the Confidence Intervals workbook. 18. A survey of 23 individuals found that they spent an
average of $39.48 on headphones to use for exercising. The margin of error for a 95% confidence interval was found to be 21.2. a. What are the lower and upper limits of the confi
dence interval? b. What was the standard deviation of the sample? 19. For the data in the Excel file Grade Point Averages,
find 90%, 95%, and 99% confidence intervals for the mean GPA. Compute the confidence intervals using the appropriate formulas and verify your results using the Excel Confidence Intervals template. 20. For the data in the Excel file Debt and Retirement
Savings, find 95% confidence intervals for the mean
income, longterm debt, and retirement savings. Use the appropriate formulas and Excel functions. 21. Find the standard deviation of the total assets held by
the bank in the Excel file Credit Risk Data. a. Treating the records in the database as a population, use your sample in Problem 2 and compute 90%, 95%, and 99% confidence intervals for the total assets held in the bank by loan applicants using formula (6.2) and any appropriate Excel functions. Explain the differences as the level of confidence increases. b. How do your confidence intervals differ if you assume that the population standard deviation is not known but estimated using your sample data? 22. A study was conducted to examine the time spent
on studying by students at a university and whether they pass or fail a course. It found that to pass or fail a course the standard deviation were 5 hours and 7 hours, respectively. The study also found that 200 students with a mean of 40.1 hours spent on studying pass a course and 245 students with a mean of 32.6 hours spent on studying fail a course. Develop 95% confidence intervals for each group of students. Use the appropriate formulas and Excel functions. 23. Using the data in the Excel file Accountants, find and
interpret 90% confidence intervals for the following: a. mean years of service b. proportion of employees who have a graduate degree Use the appropriate formulas and Excel functions. 24. The Excel file Restaurant Sales provides sample
information on lunch, dinner, and delivery sales for a local Italian restaurant. Develop 95% confidence intervals for the mean of each of these variables, as well as total sales for weekdays and weekends (Saturdays and Sundays). Use the appropriate formulas and Excel functions. 25. According to Moovit Insights website, the average
amount of time people in Singapore spend commuting to and from work, on a weekday, using public transport is 84 minutes. Recent studies have shown that 90% of people in Singapore took public transports to their workplace. A random sample of 250 individuals from Singapore is selected. Find the 90% confidence interval for the population proportion who use public transport as means of commuting.
Chapter 6 Sampling and Estimation
26. A bank sampled its customers to determine the pro
portion of customers who use their debit card at least once each month. A sample of 50 customers found that only 12 use their debit card monthly. Find 95% and 99% confidence intervals for the proportion of customers who use their debit card monthly. Use the appropriate formula and verify your result using the Confidence Intervals workbook. 27. If, based on a sample size of 850, a political candi
date finds that 458 people would vote for him in a twoperson race, what is the 95% confidence interval for his expected proportion of the vote? Would he be confident of winning based on this poll? Use the appropriate formula and verify your result using the Confidence Intervals workbook. 28. If, based on a sample size of 200, a political candi
date found that 125 people would vote for her in a twoperson race, what is the 99% confidence interval for her expected proportion of the vote? Would she be confident of winning based on this poll? Use the appropriate formula and verify your result using the Confidence Intervals workbook. 29. Using the data in the worksheet Consumer Transpor
tation Survey, develop 95% confidence intervals for the following: a. the proportion of individuals who are satisfied with their vehicle b. the proportion of individuals who have at least one child Use the appropriate formula and verify your result using the Confidence Intervals workbook. Using Confidence Intervals for Decision Making 30. A survey of 50 young professionals found that they
spent an average of $20.21 when dining out, with a standard deviation of $13.88. Can you conclude statistically that the population mean is greater than $23? 31. A survey of 270 individuals found that onesixth of
them use their cell phones primarily for email. Can you conclude statistically that the population proportion who use cell phones primarily for email is less than 0.20? 32. A manufacturer conducted a survey among 500 ran
domly selected target market households in the test market for its new disposable diapers. The objective
271
of the survey was to determine the market share for its new brand. If the sample point estimate for market share is 16%, develop a 95% confidence interval. Can the company reasonably conclude that it has a 20% market share? How about an 18% market share? 33. Using data in the Excel file Colleges and Universi
ties, find 95% confidence intervals for the median SAT for each of the two groups, liberal arts colleges and research universities. Based on these confidence intervals, does there appear to be a difference in the median SAT scores between the two groups? 34. The Excel file Professionality shows select lecturers
at a university and their related information such as gender, status of current position, number of years in current position, highest degree, number of years of working since obtaining the degree, and current salary. Develop a 99% confidence interval for the lecturers’ current salary. Interpret your findings. 35. Due to rapid developments in technology, the num
ber of laptops per household has also increased. A study was conducted for two housing areas to determine the average number and proportion of laptops per household in both areas. The data recorded by the study is as follows: Area 1 Sample size
Area 2
250
185
Average
2
3
Standard deviation
0.46
1.17
Proportion
0.57
0.63
Develop 95% confidence intervals for the average and proportion of laptops per household in both areas. 36. For the data in the Excel file Education and Income,
find 95% confidence intervals for the mean annual income of males and the mean annual income of females. Can you conclude that the mean income of one group is larger than the other? 37. For the data in the Excel file Debt and Retirement
Savings, find 95% confidence intervals for the mean income, longterm debt, and retirement savings for individuals who are single and individuals who are married. What conclusions can you reach by comparing the two groups?
272
Chapter 6 Sampling and Estimation
38. A survey of 23 patients at a hospital found 95% con
fidence intervals for the following:
Confidence Intervals and Sample Size 42. Trade associations, such as the United Dairy Industry
Association, frequently conduct surveys to identify characteristics of their membership. If this organization conducted a survey to estimate the annual per capita consumption of milk and wanted to be 99% confident that the estimate was no more than 0.92 gallon away from the actual average, what should the sample size be? Past data have indicated that the standard deviation of consumption is approximately 6 gallons.
Satisfaction with Quality of Care From Nurses
From Practitioners
Sample Proportion
0.286
0.957
Upper
1.022
0.998
Lower
0.636
0.929
Construct a stock chartbased visualization of these confidence intervals. What conclusions can you draw from the chart?
43. If a manufacturer conducted a survey among ran
domly selected target market households and wanted to be 99% confident that the difference between the sample estimate and the actual market share for its new product was no more than 6%, what sample size would be needed?
Prediction Intervals 39. Using the data in the worksheet Consumer Trans
portation Survey, develop 95% and 99% prediction intervals for the following: a. the hours per week that an individual will spend in his or her vehicle b. the number of miles driven per week
44. A gym instructor records the weights of a random
sample of 31 male high school students. The mean weight was found to be 140 pounds and the standard deviation of this population is 20 pounds. How large a sample should be selected to estimate that the mean is within 0.5 with 95% confidence level?
40. The Excel file Restaurant Sales provides sample
information on lunch, dinner, and delivery sales for a local Italian restaurant. Develop 95% prediction intervals for the daily dollar sales of each of these variables and also for the total sales dollars on a weekend day.
45. After launching a new range of health drinks, a local
fitness company conducted a survey to understand how the products were being received. From a random sample of 500 customers, who were asked to participate in the survey, 230 had bought the new products. How large a sample should be selected to estimate the population proportion is within 0.4% with 99% confidence?
41. For the Excel file Credit Approval Decisions,
find 95% prediction intervals for the credit scores and revolving balance of homeowners and nonhomeowners. How do they compare?
CASE: DROUT ADVERTISING RESEARCH PROJECT The background for this case was introduced in Chapter 2. This is a continuation of the case in Chapter 4. For this part of the case, compute confidence intervals for means and proportions and analyze the sampling errors, possibly
suggesting larger sample sizes to obtain more precise estimates. Write up your findings in a formal report or add your findings to the report you completed for the case in Chapter 4, depending on your instructor’s requirements.
Chapter 6 Sampling and Estimation
273
CASE: PERFORMANCE LAWN EQUIPMENT In reviewing your previous reports, several questions came to Elizabeth Burke’s mind. Use point and interval estimates to help answer these questions. 1. What proportion of customers rate the company with “top box” survey responses (which is defined as scale levels 4 and 5) on quality, ease of use, price, and service in the Customer Survey worksheet? How do these proportions differ by geographic region? 2. What estimates, with reasonable assurance, can PLE give customers for response times to customer service calls? 3. Engineering has collected data on alternative process costs for building transmissions in the worksheet Transmission Costs. Can you determine whether one of the proposed processes is better than the current process?
4. What would be a confidence interval for the proportion of failures of mower test performance as in the worksheet Mower Test? 5. For the data in the worksheet Blade Weight, what is the sampling distribution of the mean, the overall mean, and the standard error of the mean? Is a normal distribution an appropriate assumption for the sampling distribution of the mean? 6. How many blade weights must be measured to find a 95% confidence interval for the mean blade weight with a sampling error of at most 0.05? What if the sampling error is specified as 0.02? Answer these questions and summarize your results in a formal report to Ms. Burke.
This page intentionally left blank
CHAPTER
7
Statistical Inference
Jirsak/Shutterstock
LEARNING OBJECTIVES ■■ ■■
■■ ■■
■■ ■■
After studying this chapter, you will be able to:
Explain the purpose of hypothesis testing. Explain the difference between the null and alternative hypotheses. List the steps in the hypothesistesting procedure. State the proper forms of hypotheses for onesample hypothesis tests. Correctly formulate hypotheses. List the four possible outcome results from a hypothesis test.
■■
■■ ■■
■■
■■
Explain the difference between Type I and Type II errors. State how to increase the power of a test. Choose the proper test statistic for hypothesis tests involving means and proportions. Explain how to draw a conclusion for one and twotailed hypothesis tests. Use pvalues to draw conclusions about hypothesis tests.
276 ■■
■■
■■
Chapter 7 Statistical Inference
State the proper forms of hypotheses for twosample hypothesis tests. Select and use Excel Analysis Toolpak procedures for twosample hypothesis tests. Explain the purpose of analysis of variance.
■■
■■ ■■
Use the Excel ANOVA tool to conduct an analysis of variance test. List the assumptions of ANOVA. Conduct and interpret the results of a chisquare test for independence.
Managers need to know if the decisions they have made or are planning to make are effective. For example, they might want to answer questions like the following: Did an advertising campaign increase sales? Will product placement in a grocery store make a difference? Did a new assembly method improve productivity or quality in a factory? Many applications of business analytics involve seeking statistical evidence that decisions or process changes have met their objectives. Statistical inference focuses on drawing conclusions about populations from samples. Statistical inference includes estimation of population parameters and hypothesis testing, which is a technique that allows you to draw valid statistical conclusions about the value of population parameters or differences among them.
Hypothesis Testing Hypothesis testing involves drawing inferences about two contrasting propositions (each called a hypothesis) relating to the value of one or more population parameters, such as the mean, proportion, standard deviation, or variance. One of these propositions (called the null hypothesis) describes the existing theory or a belief that is accepted as valid unless strong statistical evidence exists to the contrary. The second proposition (called the alternative hypothesis) is the complement of the null hypothesis; it must be true if the null hypothesis is false. The null hypothesis is denoted by H0, and the alternative hypothesis is denoted by H1. Using sample data, we either 1. reject the null hypothesis and conclude that the sample data provide sufficient statistical evidence to support the alternative hypothesis, or 2. fail to reject the null hypothesis and conclude that the sample data do not support the alternative hypothesis. If we fail to reject the null hypothesis, then we can only accept as valid the existing theory or belief, but we can never prove it.
HypothesisTesting Procedure Conducting a hypothesis test involves several steps: 1. Identifying the population parameter of interest and formulating the hypotheses to test
Chapter 7 Statistical Inference
EXAMPL E 7.1
277
A Legal Analogy for Hypothesis Testing
A good analogy for hypothesis testing is the U.S. legal system. In our system of justice, a defendant is innocent until proven guilty. The null hypothesis—our belief in the absence of any contradictory evidence—is not guilty, whereas the alternative hypothesis is guilty. If the evidence (sample data) strongly indicates that the defendant is
guilty, then we reject the assumption of innocence. If the evidence is not sufficient to indicate guilt, then we cannot reject the not guilty hypothesis; however, we haven’t proven that the defendant is innocent. In reality, you can only conclude that a defendant is guilty from the evidence; you still have not proven it!
2. Selecting a level of significance, which defines the risk of drawing an incorrect conclusion when the assumed hypothesis is actually true 3. Determining a decision rule on which to base a conclusion 4. Collecting data and calculating a test statistic 5. Applying the decision rule to the test statistic and drawing a conclusion We will apply this procedure to two different types of hypothesis tests: the first involving a single population (called onesample tests) and, later, tests involving more than one population (multiplesample tests). CHECK YOUR UNDERSTANDING 1. Explain the difference between the null and alternative hypotheses. 2. List the steps in the general hypothesistesting procedure.
OneSample Hypothesis Tests A onesample hypothesis test is one that involves a single population parameter, such as the mean, proportion, standard deviation, and so on. To conduct the test, we use a single sample of data from the population. We may conduct three types of onesample hypothesis tests: H0: population parameter Ú constant vs. H1: population parameter 6 constant H0: population parameter … constant vs. H1: population parameter 7 constant H0: population parameter = constant vs. H1: population parameter ∙ constant Notice that onesample tests always compare a population parameter to some constant. For onesample tests, the statements of the null hypotheses are expressed as Ú , … , or = . It is not correct to formulate a null hypothesis using 7, 6, or ∙ . How do we determine the proper form of the null and alternative hypotheses? Hypothesis testing always assumes that H0 is true and uses sample data to determine whether H1 is more likely to be true. Statistically, we cannot “prove” that H0 is true; we can only fail to reject it. Thus, if we cannot reject the null hypothesis, we have shown only that there is insufficient evidence to conclude that the alternative hypothesis is true. However, rejecting the null hypothesis provides strong evidence (in a statistical sense) that the null hypothesis is not true and that the alternative hypothesis is true. Therefore, what we wish to provide evidence for statistically should be identified as the alternative hypothesis.
278
Chapter 7 Statistical Inference
EXAMPLE 7.2
Formulating a OneSample Test of Hypothesis
CadSoft, a producer of computeraided design software for the aerospace industry, receives numerous calls for technical support. In the past, the average response time has been at least 25 minutes. The company has upgraded its information systems and believes that this will help reduce response time. As a result, it believes that the average response time can be reduced to less than 25 minutes. The company collected a sample of 44 response times in the Excel file CadSoft Technical Support Response Times (see Figure 7.1). If the new information system makes a difference, then data should be able to confirm that the mean response
▶▶ Figure
time is less than 25 minutes; this defines the alternative hypothesis, H1. Therefore, the proper statements of the null and alternative hypotheses are H0: population mean response time Ú 25 minutes H1: population mean response time 6 25 minutes We would typically write this using the proper symbol for the population parameter. In this case, letting m be the mean response time, we would write: H0: m Ú 25 H1: m 6 25
7.1
Portion of Excel file Technical Support Response Times
Understanding Potential Errors in Hypothesis Testing We already know that sample data can show considerable variation; therefore, conclusions based on sample data may be wrong. Hypothesis testing can result in one of four different outcomes: 1. The null hypothesis is actually true, and the test correctly fails to reject it. 2. The null hypothesis is actually false, and the hypothesis test correctly reaches this conclusion. 3. The null hypothesis is actually true, but the hypothesis test incorrectly rejects it (called Type I error). 4. The null hypothesis is actually false, but the hypothesis test incorrectly fails to reject it (called Type II error). The probability of making a Type I error, that is, P(rejecting H0 ∙ H0 is true), is denoted by a and is called the level of significance. This defines the likelihood that you will make the incorrect conclusion that the alternative hypothesis is true when, in fact, the null hypothesis is true. The value of a can be controlled by the decision maker and is selected before the test is conducted. Commonly used levels for a are 0.10, 0.05, and 0.01. The probability of correctly failing to reject the null hypothesis, or P(not rejecting H0 ∙ H0 is true), is called the confidence coefficient and is calculated as 1  a. A confidence coefficient of 0.95 means that we expect 95 out of 100 samples to support the null hypothesis rather than the alternate hypothesis when H0 is actually true.
Chapter 7 Statistical Inference
279
Unfortunately, we cannot control the probability of a Type II error, P(not rejecting H0 ∙ H0 is false), which is denoted by b. Unlike a, b cannot be specified in advance but depends on the true value of the (unknown) population parameter (see Example 7.3). The value 1  b is called the power of the test and represents the probability of correctly rejecting the null hypothesis when it is indeed false, or P(rejecting H0 ∙ H0 is false). We would like the power of the test to be high (equivalently, we would like the probability of a Type II error to be low) to allow us to make a valid conclusion. The power of the test is sensitive to the sample size; small sample sizes generally result in a low value of 1  b. The power of the test can be increased by taking larger samples, which enable us to detect small differences between the sample statistics and population parameters with more accuracy. However, a larger sample size incurs higher costs, giving new meaning to the adage there is no such thing as a free lunch. This suggests that if you choose a small level of significance, you should try to compensate by having a large sample size when you conduct the test. EXAMPL E 7.3 How B Depends on the True Population Mean Consider the hypotheses in the CadSoft example: H0: mean response time Ú 25 minutes H1: mean response time 6 25 minutes If the true mean response from which the sample is drawn is, say, 15 minutes, we would expect to have a much smaller probability of incorrectly concluding that the null hypothesis is true than when the true mean response is 24 minutes, for example. If the true mean were 15 minutes, the sample mean would very likely be much less than 25, leading us to reject H0. If the true mean were 24 minutes,
even though it is less than 25, we would have a much higher probability of failing to reject H0 because a higher likelihood exists that the sample mean would be greater than 25 due to sampling error. Thus, the farther away the true mean response time is from the hypothesized value, the smaller is b. Generally, as a decreases, b increases, so the decision maker must consider the tradeoffs of these risks. So if you choose a level of significance of 0.01 instead of 0.05 and keep the sample size constant, you would reduce the probability of a Type I error but increase the probability of a Type II error.
Selecting the Test Statistic The next step is to collect sample data and use the data to draw a conclusion. The decision to reject or fail to reject a null hypothesis is based on computing a test statistic from the sample data. The test statistic used depends on the type of hypothesis test. Different types of hypothesis tests use different test statistics, and it is important to use the correct one. The proper test statistic often depends on certain assumptions about the population—for example, whether or not the standard deviation is known. In the vast majority of practical applications, the population standard deviation is unknown. The following formulas show two types of onesample hypothesis tests for means and their associated test statistics. The value of m0 is the hypothesized value of the population mean, that is, the “constant” in the hypothesis formulation. Type of Test
Test Statistic
Onesample test for mean, s known
z =
Onesample test for mean, s unknown
t=
x  m0 s> 1n
x  m0 s> 1n
(7.1) (7.2)
With sample data, we generally do not know the population standard deviation, and therefore we only will illustrate an example where s is unknown. We will summarize the procedure when s is known later in this chapter.
280
Chapter 7 Statistical Inference
EXAMPL E 7.4
Computing the Test Statistic
For the CadSoft example, the average response time for the sample of 44 customers is x = 21.91 minutes and the sample standard deviation is s = 19.49. The hypothesized mean is m 0 = 25. You might wonder why we have to test the hypothesis statistically when the sample average of 21.91 is clearly less than 25. We do because of sampling error. It is quite possible that the population mean truly is 25 or more and that we were lucky to draw a sample whose mean was smaller. Because of potential sampling error, it would be dangerous to conclude that the company was meeting its goal just by looking at the sample mean without better statistical evidence. Because we don’t know the value of the population standard deviation, the proper test statistic to use is formula (7.2): t =
x  m0
Therefore, the value of the test statistic is t =
21.91  25 3.09 x  m0 = = = 1.05 2.938 s> 1n 19.49> 144
Observe that the numerator is the distance between the sample mean (21.91) and the hypothesized value (25). By dividing by the standard error, the value of t represents the number of standard errors the sample mean is from the hypothesized value. In this case, the sample mean is 1.05 standard errors below the hypothesized value of 25. This notion provides the fundamental basis for the hypothesis test—if the sample mean is “too far” away from the hypothesized value, then the null hypothesis should be rejected.
s> 1n
Finding Critical Values and Drawing a Conclusion The conclusion to reject or fail to reject H0 is based on comparing the value of the test statistic to a “critical value” from the sampling distribution of the test statistic when the null hypothesis is true and the chosen level of significance, a. The sampling distribution of the test statistic is usually the normal distribution, tdistribution, or some other wellknown distribution. For example, the sampling distribution of the ztest statistic in formula (7.1) is a standard normal distribution; the ttest statistic in formula (7.2) has a tdistribution with n  1 degrees of freedom. For a onetailed test, the critical value is the number of standard errors away from the hypothesized value for which the probability of exceeding the critical value is a. If a = 0.05, for example, then we are saying that there is only a 5% chance that a sample mean will be that far away from the hypothesized value purely because of sampling error. Should this occur, it suggests that the true population mean is different from what was hypothesized. The critical value divides the sampling distribution into two parts, a rejection region and a nonrejection region. If the null hypothesis is false, it is more likely that the test statistic will fall into the rejection region. If it does, we reject the null hypothesis; otherwise, we fail to reject it. The rejection region is chosen so that the probability of the test statistic falling into it if H0 is true is the probability of a Type I error, a. The rejection region occurs in the tails of the sampling distribution of the test statistic and depends on the structure of the hypothesis test, as shown in Figure 7.2. If the null hypothesis is structured as = and the alternative hypothesis as ∙, then we would reject H0 if the test statistic is either significantly high or low. In this case, the rejection region will occur in both the upper and lower tail of the distribution [see Figure 7.2(a)]. This is called a twotailed test of hypothesis. Because the probability that the test statistic falls into the rejection region, given that H0 is true, is a, each tail has an area of a>2.
▶▶ Figure
281
Chapter 7 Statistical Inference
7.2
Rejection Region
Illustration of Rejection Regions in Hypothesis Testing
c/2
c/2
Lower critical value Upper critical value (a) Twotailed test
Rejection Region
Rejection Region
c
c Lower onetailed test
Critical value
Upper onetailed test (b) Onetailed tests
Critical value
The other types of hypothesis tests, which specify a direction of relationship (where H0 is either Ú or … ), are called onetailed tests of hypothesis. In this case, the rejection region occurs only in one tail of the distribution [see Figure 7.2(b)]. Determining the correct tail of the distribution to use as the rejection region for a onetailed test is easy. If H1 is stated as 6 , the rejection region is in the lower tail; if H1 is stated as 7 , the rejection region is in the upper tail (just think of the inequality associated with the alternative hypothesis as an arrow pointing to the proper tail direction). Twotailed tests have both upper and lower critical values, whereas onetailed tests have either a lower or an upper critical value. For standard normal and tdistributions, which have a mean of zero, lowertail critical values are negative; uppertail critical values are positive. Critical values make it easy to determine whether or not the test statistic falls in the rejection region of the proper sampling distribution. For example, for an upper onetailed test, if the test statistic is greater than the critical value, the decision would be to reject the null hypothesis. Similarly, for a lower onetailed test, if the test statistic is less than the critical value, we would reject the null hypothesis. For a twotailed test, if the test statistic is either greater than the upper critical value or less than the lower critical value, the decision would be to reject the null hypothesis. The critical value for a onesample, onetailed test when the standard deviation is unknown is the value of the tdistribution with n  1 degrees of freedom that provides a tail area of alpha, that is, ta,n  1. We may find tvalues in Table A.2 in Appendix A at the end of the book or by using the Excel function T.INV(1  a, n  1). Again, for a lowertailed test, the critical value is negative, and for a twotailed test, we would use a>2.
282
Chapter 7 Statistical Inference
EXAMPLE 7.5
Finding the Critical Value and Drawing a Conclusion
For the CadSoft example, if the level of significance is 0.05, then the critical value is t0.05,43 is found in Excel using = T.INV10.95, 432 = 1.68. Because the tdistribution is symmetric with a mean of 0 and this is a lowertailed test, we use the negative of this number ( 1.68) as the critical value. By comparing the value of the ttest statistic with this critical value, we see that the test statistic computed in Example 7.4 does not fall below the critical value (that is,
▶▶ Figure
1.05 7 1.68) and is not in the rejection region. Therefore, we cannot reject H0 and cannot conclude that the mean response time has improved to less than 25 minutes. Figure 7.3 illustrates the conclusion we reached. Even though the sample mean is less than 25, we cannot conclude that the population mean response time is less than 25 because of the large amount of sampling error.
7.3
tTest for Mean Response Time
Rejection Region
t 1.68 1.05
0
TwoTailed Test of Hypothesis for the Mean Basically, all hypothesis tests are similar; you just have to ensure that you select the correct test statistic, critical value, and rejection region, depending on the type of hypothesis. The following example illustrates a twotailed test of hypothesis for the mean. For a twotailed test using the tdistribution, we use the Excel function T.INV(1  a>2, n  1) or T.INV.2T(a, n  1) to find the critical value, ta>2,n1. Be careful when using T.INV.2T! Use a and not a>2 in this function. EXAMPLE 7.6
Conducting a TwoTailed Hypothesis Test for the Mean
Figure 7.4 shows a portion of data collected in a survey of 34 respondents by a travel agency (provided in the Excel file Vacation Survey). Suppose that the travel agency wanted to target individuals who were approximately 35 years old. Thus, we wish to test whether the average age of respondents is equal to 35. The hypothesis to test is H0: mean age = 35 H1: mean age ∙ 35 The sample mean is computed to be 38.676, and the sample standard deviation is 7.857. We use the ttest statistic: t =
x  m0 s> 1n
=
38.676  35 .857> 234
= 2.73
In this case, the sample mean is 2.73 standard errors above the hypothesized mean of 35. However, because this is a twotailed test, the rejection region and decision rule are different. For a level of significance a, we reject H0 if the ttest statistic falls either below the negative critical value, ta>2,n1, or above the positive critical value, ta>2,n1. Using either Table A.2 in Appendix A at the back of this book or the Excel function =T.INV.2T(.05, 33) to calculate t0.025,33, we obtain 2.0345. Thus, the critical values are { 2.0345. Because the ttest statistic does not fall between these values, we must reject the null hypothesis that the average age is 35 (see Figure 7.5).
▶▶ Figure
Chapter 7 Statistical Inference
283
7.4
Portion of Vacation Survey Data
▶▶ Figure
7.5
Rejection Region
Illustration of a TwoTailed Test for Example 7.6
– 2.0345
0
2.0345 2.73
Summary of OneSample Hypothesis Tests for the Mean Case 1: (s Unknown) 1. Determine whether the proper hypotheses represent a lowertailed, uppertailed, or twotailed test. 2. Calculate the test statistic using formula (7.2). 3. Find the critical value. a. If it is a lowertailed test, the critical value is found using the Excel function T.INV(1  a, n  1). Note the minus sign! b. If it is an uppertailed test, the critical value is found using the Excel f unction T.INV(1  a, n  1). c. If you have a twotailed test, use T.INV.2T(a, n  1); the critical values will be both positive and negative. 4. Compare the test statistic to the critical value(s) and draw a conclusion to either reject the null hypothesis or fail to reject it. When the population standard deviation is known, the process is exactly the same; only the formulas and Excel functions differ. Case 2: (s Known) 1. Determine whether the proper hypotheses represent a lowertailed, uppertailed, or twotailed test. 2. Calculate the test statistic using formula (7.1). 3. Find the critical value. The critical value for a onesample, onetailed test when the standard deviation is known is the value of the normal distribution that has a tail area of alpha. This may be found by using Table A.1 in Appendix A to find the zvalue corresponding to an area of 1  a or using the Excel function
284
Chapter 7 Statistical Inference
NORM.S.INV(1  a). Remember that for a lowertailed test, the critical value is negative. For a twotailed test, use a>2. a. If it is a lowertailed test, the critical value is found using the Excel function NORM.S.INV(1  a). Note the minus sign! b. If it is an uppertailed test, the critical value is found using the Excel function NORM.S.INV(1  a). c. If you have a twotailed test, use NORM.S.INV(1  a>2); the critical values will be both positive and negative. 4. Compare the test statistic to the critical value(s) and draw a conclusion to either reject the null hypothesis or fail to reject it.
pValues An alternative approach to comparing a test statistic to a critical value in hypothesis testing is to find the probability of obtaining a test statistic value equal to or more extreme than that obtained from the sample data when the null hypothesis is true. This probability is commonly called a pvalue, or observed significance level. The magnitude of a pvalue indicates the compatibility or incompatibility of the data with the null hypothesis. The smaller the pvalue, the greater the statistical incompatibility with the null hypothesis. The pvalue does not measure the probability that the null hypothesis is true! It is common practice to compare pvalues with the level of significance chosen for a hypothesis test. To draw a conclusion, compare the pvalue to the chosen level of significance a; whenever p 6 a, reject the null hypothesis and otherwise fail to reject it. Using pvalues makes it easy to draw conclusions about hypothesis tests. If the population standard deviation is known, the pvalue for a lower onetailed test is the probability to the left of the test statistic z and is found by =NORM.S.DIST(z, TRUE). For an upper onetailed test, the pvalue is the probability to the right of the test statistic z and is found by =1NORM.S.DIST(z, TRUE). For a twotailed test, the pvalue is the probability to the left of the negative zvalue plus the probability to the right of the positive zvalue and can be computed in Excel as = 2*(1  NORM.S.DIST(ABS(z), TRUE). When the population standard deviation is unknown, the pvalue for a lower onetailed test is the probability to the left of the test statistic t in the tdistribution and is found by =T.DIST(t, n  1, TRUE). For an upper onetailed test, the pvalue is the probability to the right of the test statistic t and is found by =1 T.DIST(t, n  1, TRUE). For a twotailed test, the pvalue can be found using the Excel function =T.DIST.2T(ABS(t), n  1). The statistical community has questioned this practice. The American Statistical Association has noted1 Practices that reduce data analysis or scientific inference to mechanical “brightline” rules (such as “p 6 0.05”) for justifying scientific claims or conclusions can lead to erroneous beliefs and poor decisionmaking. A conclusion does not immediately become “true” on one side of the divide and “false” on the other. Researchers should bring many contextual factors into play to derive scientific inferences, including the design of a study, the quality of the measurements, the external evidence for the phenomenon under study, and the validity of assumptions that underlie the data analysis. Pragmatic considerations often require binary, “yesno” decisions, but this does not mean that pvalues alone can ensure that a decision is correct or incorrect. The widespread use of “statistical significance” (generally interpreted as “p … 0.05”) as a license for making a claim of a scientific finding (or implied truth) leads to considerable distortion of the scientific process. As a result, pvalues should be interpreted cautiously. 1
Ronald L. Wasserstein and Nicole A. Lazar (2016), “The ASA’s statement on pvalues: context, process, and purpose,” The American Statistician, DOI: 10.1080/00031305.2016.1154108
Chapter 7 Statistical Inference
285
EXAMPL E 7.7 Using pValues For the CadSoft example (see Example 7.4), the ttest statistic for the hypothesis test in the responsetime example is 1.05. If the true mean is really 25, then the pvalue is the probability of obtaining a test statistic of 1.05 or less (the area to the left of 1.05 in Figure 7.3). We can calculate the pvalue using the Excel function = T.DIST1 1.05, 43, TRUE2 = 0.15 . Because p = 0.15 is not less than a = 0.05, we do not reject H0. In other words, there is about a 15% chance that the test statistic would
be 1.05 or smaller if the null hypothesis were true. This is a fairly high probability, so it would be difficult to conclude that the true mean is less than 25 and we could attribute the fact that the test statistic is less than the hypothesized value to sampling error alone and not reject the null hypothesis. For the Vacation Survey twotailed hypothesis test in Example 7.6, the pvalue for this test is 0.010, which can also be computed by the Excel function =T.DIST.2T(2.73, 33); therefore, since 0.010 6 0.05, we reject H0.
OneSample Tests for Proportions Many important business measures, such as market share or the fraction of deliveries received on time, are expressed as proportions. We may conduct a test of hypothesis about a population proportion in a similar fashion as we did for means. The test statistic for a onesample test for proportions is
z =
pn  p0 2p0 11  p0 2 >n
(7.3)
where p0 is the hypothesized value and pn is the sample proportion. Similar to the test statistic for means, the ztest statistic shows the number of standard errors that the sample proportion is from the hypothesized value. The sampling distribution of this test statistic has a standard normal distribution. For a lowertailed test, the pvalue would be computed by the area to the left of the test statistic; that is, =NORM.S.DIST(z, TRUE). For an uppertailed test, the pvalue would be the area to the right of the test statistic, or = 1 NORM.S.DIST(z, TRUE). If we had a twotailed test, the pvalue is computed by the Excel formula =2*(1NORM.S.DIST (ABS(z), TRUE)). EXAMPL E 7.8
A OneSample Test for the Proportion
CadSoft also sampled 44 customers and asked them to rate the overall quality of the company’s software product using the following scale: 0—very poor 1—poor 2—good 3—very good 4—excellent These data can be found in the Excel File CadSoft Product Satisfaction Survey. The firm tracks customer satisfaction of quality by measuring the proportion of responses in the top two categories. In the past, this proportion has averaged about 75%. For these data, 35 of the 44 responses, or 79.5%, are in the top two categories. Is there sufficient evidence to conclude that this satisfaction measure has significantly exceeded 75%
using a significance level of 0.05? Answering this question involves testing the hypotheses about the population proportion p: H0: p … 0.75 H1: p 7 0.75 This is an uppertailed, onetailed test. The test statistic is computed using formula (7.3): z =
0.795  0.75 20.75(1  0.75)>44
= 0.69
In this case, the sample proportion of 0.795 is 0.69 standard error above the hypothesized value of 0.75. Because this is an uppertailed test, we reject H0 if the value of the test statistic is larger than the critical value. Because the sampling distribution of z is a standard normal, the critical
(continued )
286
Chapter 7 Statistical Inference
value of z for a level of significance of 0.05 is found by the Excel function NORM.S. INV10.952 = 1.645. Because the test statistic does not exceed the critical value, we cannot reject the null hypothesis that the proportion is no greater than 0.75. Thus, even though the sample proportion exceeds 0.75, we cannot conclude statistically that the customer satisfaction ratings have significantly
improved. We could attribute this to sampling error and the relatively small sample size. The pvalue can be found by computing the area to the right of the test statistic in the standard normal distribution using the Excel formula = 1  NORM.S.DIST(0.69, TRUE) = 0.25. Note that the pvalue is greater than the significance level of 0.05, leading to the same conclusion of not rejecting the null hypothesis.
Confidence Intervals and Hypothesis Tests A close relationship exists between confidence intervals and hypothesis tests. For example, suppose we construct a 95% confidence interval for the mean. If we wish to test the hypotheses H0: m = m0 H1: m ∙ m0 at a 5% level of significance, we simply check whether the hypothesized value m0 falls within the confidence interval. If it does not, then we reject H0; if it does, then we cannot reject H0. For onetailed tests, we need to examine on which side of the hypothesized value the confidence interval falls. For a lowertailed test, if the confidence interval falls entirely below the hypothesized value, we reject the null hypothesis. For an uppertailed test, if the confidence interval falls entirely above the hypothesized value, we also reject the null hypothesis.
An Excel Template for OneSample Hypothesis Tests The Excel file One Sample Hypothesis Tests provides template worksheets for conducting hypothesis tests for means (with s known and unknown) and proportions based on the formulas and Excel functions that we have introduced in this section. The templates provide results for upper and lower onetailed tests, and also for a twotailed test of hypothesis. Figure 7.6 shows the templates for the CadSoft examples. We strongly suggest that you examine the Excel formulas and compare them to the examples and rules that we have presented to better understand the templates.
CHECK YOUR UNDERSTANDING 1. State the three types of onesample hypothesis tests. 2. Why is it important to identify the alternative hypothesis based on what we wish to provide statistical evidence for? 3. Explain the concepts of Type I and Type II errors. 4. Explain the difference between a onetailed and twotailed hypothesis test. 5. How do you use a test statistic to draw a conclusion about a hypothesis test? 6. What is a pvalue, and how is it used?
Chapter 7 Statistical Inference
▶▶ Figure
287
7.6
Excel Templates for OneSample Hypothesis Tests
TwoSample Hypothesis Tests Many practical applications of hypothesis testing involve comparing two populations for differences in means, proportions, or other population parameters. Such tests can confirm differences between suppliers, performance at two factory locations, new and old work methods or reward and recognition programs, and many other situations. Similar
288
Chapter 7 Statistical Inference
to onesample tests, twosample hypothesis tests for differences in population parameters have one of the following forms: 1. Lowertailed test H0: population parameter (1)  population parameter (2) Ú D0 vs. H1: population parameter (1)  population parameter (2) 6 D0. This test seeks evidence that the difference between population parameter (1) and population parameter (2) is less than some value, D0. When D0 = 0, the test simply seeks to conclude whether population parameter (1) is smaller than population parameter (2). 2. Uppertailed test H0: population parameter (1)  population parameter (2) … D0 vs. H1: population parameter (1)  population parameter (2) 7 D0. This test seeks evidence that the difference between population parameter (1) and population parameter (2) is greater than some value, D0. When D0 = 0, the test simply seeks to conclude whether population parameter (1) is larger than population parameter (2). 3. Twotailed test H0: population parameter (1)  population parameter (2) = D0 vs. H1: population parameter (1)  population parameter (2) ∙ D0. This test seeks evidence that the difference between the population parameters is equal to D0. When D0 = 0, we are seeking evidence that population parameter (1) differs from parameter (2). In most applications, D0 = 0, and we are simply seeking to compare the population parameters. However, there are situations when we might want to determine if the parameters differ by some nonzero amount; for example, “job classification A makes at least $5,000 more than job classification B.” These hypothesistesting procedures are similar to those previously discussed in the sense of computing a test statistic and comparing it to a critical value. However, the test statistics for twosample tests are more complicated than those for onesample tests and we will not delve into the mathematical details. Fortunately, Excel provides several tools for conducting twosample tests, and we will use these in our examples. Table 7.1 summarizes the Excel Analysis Toolpak procedures that we will use.
TwoSample Tests for Differences in Means In a twosample test for differences in means, we always test hypotheses of the form H0: m1  m2 5 Ú , … , or = 6 0
H1: m1  m2 5 6 , 7 , or ∙ 6 0
▶▶ Table
7.1
Excel Analysis Toolpak Procedures for TwoSample Hypothesis Tests
(7.4)
Type of Test
Excel Procedure
Twosample test for means, s2 known
Excel ztest: Twosample for means
Twosample test for means, s2 unknown and assumed unequal
Excel ttest: Twosample assuming unequal variances
Twosample test for means, s2 unknown and assumed equal
Excel ttest: Twosample assuming equal variances
Paired twosample test for means
Excel ttest: Paired twosample for means
Twosample test for equality of variances
Excel Ftest: Twosample for variances
Chapter 7 Statistical Inference
EXAMPL E 7.9
289
Comparing Supplier Performance
The last two columns in the Purchase Orders data file provide the order date and arrival date of all orders placed with each supplier. The time between placement of an order and its arrival is commonly called the lead time. We may compute the lead time by subtracting the Excel date function values from each other (Arrival Date  Order Date), as shown in Figure 7.7. Figure 7.8 shows a PivotTable for the average lead time for each supplier. Purchasing managers have noted that they order many of the same types of items from Alum Sheeting and Durrable Products and are considering dropping Alum Sheeting from their supplier base if its lead time is significantly longer than that of Durrable Products. Therefore, they would like to test the hypothesis
H0: m1  m2 … 0 H1: m1  m2 7 0 where m1 = mean lead time for Alum Sheeting and m2 = mean lead time for Durrable Products. Rejecting the null hypothesis suggests that the average lead time for Alum Sheeting is statistically longer than Durrable Products. However, if we cannot reject the null hypothesis, then even though the mean lead time for Alum Sheeting is longer, the difference would most likely be due to sampling error, and we could not conclude that there is a statistically significant difference.
Selection of the proper test statistic and Excel procedure for a twosample test for means depends on whether the population variances are known, and if not, whether they are assumed to be equal. 1. Population variance is known. In Excel, choose zTest: TwoSample for Means from the Data Analysis menu. This test uses a test statistic that is based on the standard normal distribution. 2. Population variance is unknown and assumed unequal. From the Data Analysis menu, choose tTest: TwoSample Assuming Unequal Variances. The test statistic for this case has a tdistribution. 3. Population variance is unknown but assumed equal. In Excel, choose tTest: TwoSample Assuming Equal Variances. The test statistic also has a tdistribution, but it is different from the unequal variance case. These tools calculate the test statistic, the pvalue for both a onetailed and twotailed test, and the critical values for onetailed and twotailed tests. For the ztest with known population variances, these are called z, P1Z … z2 onetail or P1Z … z2 twotail, and z Critical onetail or z Critical twotail, respectively. For the ttests, these are called t Stat, P1T … t2 onetail or P1T … t2 twotail, and t Critical onetail or t Critical twotail, respectively. Caution: You must be very careful in interpreting the output information from these Excel tools and apply the following rules: ▼ Figure
7.7
Portion of Purchase Orders Database with Lead Time Calculations
1. If the test statistic is negative, the onetailed pvalue is the correct pvalue for a lowertailed test; however, for an uppertailed test, you must subtract this number from 1.0 to get the correct pvalue.
290
Chapter 7 Statistical Inference
▶▶ Figure
7.8
Pivot Table for Average Supplier Lead Time
2. If the test statistic is nonnegative (positive or zero), then the pvalue in the output is the correct pvalue for an uppertailed test; but for a lowertailed test, you must subtract this number from 1.0 to get the correct pvalue. 3. For a lowertailed test, you must change the sign of the onetailed critical value. Only rarely are the population variances known; also, it is often difficult to justify the assumption that the variances of each population are equal. Therefore, in most practical situations, we use the tTest: TwoSample Assuming Unequal Variances. This procedure also works well with small sample sizes if the populations are approximately normal. It is recommended that the size of each sample be approximately the same and total 20 or more. If the populations are highly skewed, then larger sample sizes are recommended. EXAMPLE 7.10 Testing the Hypotheses for Supplier LeadTime Performance To conduct the hypothesis test for comparing the lead times for Alum Sheeting and Durrable Products, first sort the data by supplier and then select tTest: TwoSample Assuming Unequal Variances from the Data Analysis menu. The dialog is shown in Figure 7.9. The dialog prompts you for the range of the data for each variable, hypothesized mean difference, whether the ranges have labels, and the level of significance a. If you leave the box Hypothesized Mean Difference blank or enter zero, the test is for equality of means. However, the tool allows you to specify a value D0 to test the hypothesis H0: m1  m2 = D0 if you want to test whether the population means have a certain distance between them. In this example, the Variable 1 range defines the lead times for Alum Sheeting and the Variable 2 range for Durrable Products.
Figure 7.10 shows the results from the tool. The tool provides information for both onetailed and twotailed tests. Because this is a onetailed test, we use the highlighted information in Figure 7.10 to draw our conclusions. For this example, t Stat is positive and we have an uppertailed test; therefore, using the rules stated earlier, the pvalue is 0.00166. Based on this alone, we reject the null hypothesis and must conclude that Alum Sheeting has a statistically longer average lead time than Durrable Products. We may draw the same conclusion by comparing the value of t Stat with the critical value t Critical onetail. Being an uppertailed test, the value of t Critical onetail is 1.812. Comparing this with the value of t Stat, we would reject H0 only if t Stat 7 t Critical [email protected] Since t Stat is greater than t Critical onetail, we reject the null hypothesis.
TwoSample Test for Means with Paired Samples In the previous example for testing differences in the mean supplier lead times, we used independent samples; that is, the orders in each supplier’s sample were not related to each other. In many situations, data from two samples are naturally paired or matched. For example, suppose that a sample of assembly line workers perform a task using two different types of work methods, and the plant manager wants to determine if any differences exist between the two methods. In collecting the data, each worker will have performed
▶▶ Figure
Chapter 7 Statistical Inference
291
7.9
Dialog for TwoSample tTest, Sigma Unknown
▶▶ Figure
7.10
Results for TwoSample Test for LeadTime Performance
the task using each method. Had we used independent samples, we would have randomly selected two different groups of employees and assigned one work method to one group and the alternative method to the second group. Each worker would have performed the task using only one of the methods. As another example, suppose that we wish to compare retail prices of grocery items between two competing grocery stores. It makes little sense to compare different samples of items from each store. Instead, we would select a sample of grocery items and find the price charged for the same items by each store. In this case, the samples are paired because each item would have a price from each of the two stores. When paired samples are used, a paired ttest is more accurate than assuming that the data come from independent populations. The null hypothesis we test revolves around the mean difference (mD) between the paired samples; that is, H0: mD 5 Ú , … , or = 6 0
H1: mD 5 6 , 7 , or ∙ 6 0
The test uses the average difference between the paired data and the standard deviation of the differences similar to a onesample test. Excel has a Data Analysis tool, tTest: Paired TwoSample for Means, for conducting this type of test. In the dialog, you need to enter only the variable ranges and hypothesized mean difference.
292
Chapter 7 Statistical Inference
EXAMPLE 7.11
Using the Paired TwoSample Test for Means
The Excel file Pile Foundation contains the estimates used in a bid and actual augercast pile lengths that engineers ultimately had to use for a foundationengineering project. The contractor’s past experience suggested that the bid information was generally accurate, so the average of the paired differences between the actual pile lengths and estimated lengths should be close to zero. After this project was completed, the contractor found that the average difference between the actual lengths and the estimated lengths was 6.38. Could the contractor conclude that the bid information was poor? Figure 7.11 shows a portion of the data and the Excel dialog for the paired twosample test. Figure 7.12
shows the output from the Excel tool using a significance level of 0.05, where Variable 1 is the estimated lengths, and Variable 2 is the actual lengths. This is a twotailed test, so in Figure 7.12 we interpret the results using only the twotail information that is highlighted. The critical values are { 1.968, and because t Stat is much smaller than the lower critical value, we must reject the null hypothesis and conclude that the mean of the differences between the estimates and the actual pile lengths is statistically significant. Note that the pvalue is essentially zero, verifying this conclusion.
TwoSample Test for Equality of Variances Understanding variation in business processes is very important, as we have stated before. For instance, does one location or group of employees show higher variability than others? We can test for equality of variances between two samples using a new type of test, the Ftest. To use this test, we must assume that both samples are drawn from normal populations. The hypotheses we test are ▶▶ Figure
7.11
Portion of Excel File Pile Foundation
▶▶ Figure
7.12
Excel Output for Paired TwoSample Test for Means
Chapter 7 Statistical Inference
H0: s21  s22 = 0
H1: s21  s22 ∙ 0
293
(7.5)
To test these hypotheses, we collect samples of n1 observations from population 1 and n2 observations from population 2. The test uses an Ftest statistic, which is the ratio of the variances of the two samples: s21
(7.6) s22 Because the test statistic is a ratio of variances, an alternate way of expressing the hypotheses that some statistians prefer is
F =
H0: H1:
s21 s22 s21 s22
= 1 ∙ 1
However, this form is equivalent to (7.5). The sampling distribution of this statistic is called the Fdistribution. Similar to the tdistribution, it is characterized by degrees of freedom; however, the Fdistribution has two degrees of freedom, one associated with the numerator of the Fstatistic, n1  1, and one associated with the denominator of the Fstatistic, n2  1. Table A.4 in Appendix A at the end of the book provides only uppertail critical values, and the distribution is not symmetric, as is the standard normal distribution or the tdistribution. Therefore, although the hypothesis test is really a twotailed test, we will simplify it as a onetailed test to make it easy to use tables of the Fdistribution and interpret the results of the Excel tool that we will use. We do this by ensuring that when we compute F, we take the ratio of the larger sample variance to the smaller sample variance. EXAMPL E 7.12 Applying the FTest for Equality of Variances To illustrate the Ftest, suppose that we wish to determine whether the variance of lead times is the same for Alum Sheeting and Durrable Products in the Purchase Orders data. The Ftest can be applied using the Excel Data Analysis tool Ftest TwoSample for Variances. The dialog prompts you to enter the range of the sample data for each variable. As we noted, you should ensure that the first variable has the larger variance; this might require you to calculate the variances before you use the tool. In this case, the variance of the lead times for Alum Sheeting is larger than the variance for Durrable Products (see Figure 7.10), ▶▶ Figure
7.13
Results for TwoSample FTest for Equality of Variances
so this is assigned to Variable 1. Note also that if we choose a = 0.05, we must enter 0.025 for the level of significance in the Excel dialog. The results are shown in Figure 7.13. The value of the Fstatistic, F, is 3.467. We compare this with the uppertailed critical value, F Critical onetail, which is 3.607. Because F 6 F Critical onetail, we cannot reject the null hypothesis and conclude that the variances are not significantly different from each other. Note that the pvalue is P(F6 = f ) one tail = 0.0286. Although the level of significance is 0.05, remember that we must compare this to a> 2 = 0.025 because we are using only uppertailed information.
294
Chapter 7 Statistical Inference
If the variances differ significantly from each other, we would expect F to be much larger than 1; the closer F is to 1, the more likely it is that the variances are the same. Therefore, we need only to compare F to the uppertailed critical value. Hence, for a level of significance a, we find the critical value Fa>2,df1,df2 of the Fdistribution, and then we reject the null hypothesis if the Ftest statistic exceeds the critical value. Note that we are using a>2 to find the critical value, not a. This is because we are using only the uppertailed information on which to base our conclusion. The Ftest for equality of variances is often used before testing for the difference in means so that the proper test (population variance is unknown and assumed unequal or population variance is unknown and assumed equal, which we discussed earlier in this chapter) is selected. CHECK YOUR UNDERSTANDING 1. State the only correct forms of twosample hypothesis tests. 2. What Excel procedures are available for conducting twosample hypothesis tests? 3. How does a paired sample test for means differ from other twosample hypothesis tests for means? 4. Explain how to conduct a test for equality of variances.
Analysis of Variance (ANOVA) To this point, we have discussed hypothesis tests that compare a population parameter to a constant value or that compare the means of two different populations. Often, we would like to compare the means of more than two groups to determine if all are equal or if any are significantly different from the rest. EXAMPLE 7.13
Differences in Insurance Survey Data
In the Excel data file Insurance Survey, we might be interested in whether any significant differences exist in satisfaction among individuals with different levels of education. We could sort the data by educational level and then create a table similar to the one shown below.
College Graduate
Although the average satisfaction for each group is somewhat different and it appears that the mean satisfaction of individuals with a graduate degree is higher, we cannot tell conclusively whether or not these differences are significant because of sampling error.
Graduate Degree
Some College
5
3
4
3
4
1
5
5
4
3
5
2
3
5
3
3
4
4
3
5
4
4
5
2 Average
3.444
4.500
3.143
Count
9
8
7
Chapter 7 Statistical Inference
295
In statistical terminology, the variable of interest is called a factor. In this example, the factor is the educational level, and we have three categorical levels of this factor, college graduate, graduate degree, and some college. Thus, it would appear that we will have to perform three different pairwise tests to establish whether any significant differences exist among them. As the number of factor levels increases, you can easily see that the number of pairwise tests grows large very quickly. Fortunately, other statistical tools exist that eliminate the need for such a tedious approach. Analysis of variance (ANOVA) is one of them. Suppose we have m groups. The null hypothesis for ANOVA is that the population means of all m groups are equal; the alternative hypothesis is that at least one mean differs from the rest: H0: m1 = m2 = g = mm H1: at least one mean is different from the others ANOVA derives its name from the fact that we are analyzing variances in the data; essentially, ANOVA computes a measure of the variance between the means of each group and a measure of the variance within the groups and examines a test statistic that is the ratio of these measures. This test statistic can be shown to have an Fdistribution (similar to the test for equality of variances). If the Fstatistic is large enough based on the level of significance chosen and exceeds a critical value, we would reject the null hypothesis. Excel provides a Data Analysis tool, ANOVA: Single Factor, to conduct analysis of variance. EXAMPL E 7.14
Applying the Excel ANOVA Tool
To test the null hypothesis that the mean satisfaction for all educational levels in the Excel file Insurance Survey are equal against the alternative hypothesis that at least one mean is different, select ANOVA: Single Factor from the Data Analysis options. First, you must set up the worksheet so that the data you wish to use are displayed in contiguous columns, as shown in Example 7.13. In the dialog shown in Figure 7.14, specify the input range of the data (which must be in contiguous columns) and whether it is stored in rows or columns (that is, whether each factor level or group is a row or column in the range). The sample size for each factor level need not be the same, but the input range must be a rectangular region that contains all data. You must also specify the level of significance (a) .
▶▶ Figure
7.14
ANOVA: Single Factor Dialog
The results for this example are given in Figure 7.15. The output report begins with a summary report of basic statistics for each group. The ANOVA section reports the details of the hypothesis test. You needn’t worry about all the mathematical details. The important information to interpret the test is given in the columns labeled F (the Ftest statistic), Pvalue (the pvalue for the test), and F crit (the critical value from the Fdistribution). In this example, F = 3.92, and the critical value from the Fdistribution is 3.4668. Here F 7 F crit; therefore, we must reject the null hypothesis and conclude that there are significant differences in the means of the groups—that is, the mean satisfaction is not the same among the three educational levels. Alternatively, we see that the pvalue is smaller than the chosen level of significance, 0.05, leading to the same conclusion.
296
Chapter 7 Statistical Inference
▶▶ Figure
7.15
ANOVA Results for Insurance Survey Data
Although ANOVA can identify a difference among the means of multiple populations, it cannot determine which means are different from the rest. To do this, we may use the TukeyKramer multiple comparison procedure. Unfortunately, Excel does not provide this tool, but it may be found in other statistical software.
Assumptions of ANOVA ANOVA requires assumptions that the m groups or factor levels being studied represent populations whose outcome measures 1. are randomly and independently obtained, 2. are normally distributed, and 3. have equal variances. If these assumptions are violated, then the level of significance and the power of the test can be affected. Usually, the first assumption is easily validated when random samples are chosen for the data. ANOVA is fairly robust to departures from normality, so in most cases this isn’t a serious issue. If sample sizes are equal, violation of the third assumption does not have serious effects on the statistical conclusions; however, with unequal sample sizes, it can. When the assumptions underlying ANOVA are violated, you may use a nonparametric test that does not require these assumptions; we refer you to more comprehensive texts on statistics for further information and examples. Finally, we wish to point out that students often use ANOVA to compare the equality of means of exactly two populations. It is important to realize that by doing this, you are making the assumption that the populations have equal variances (assumption 3). Thus, you will find that the pvalues for both ANOVA and the tTest: TwoSample Assuming Equal Variances will be the same and lead to the same conclusion. However, if the variances are unequal, as is generally the case with sample data, ANOVA may lead to an erroneous conclusion. We recommend that you do not use ANOVA for comparing the means of two populations, but instead use the appropriate ttest that assumes unequal variances. CHECK YOUR UNDERSTANDING 1. What hypotheses does ANOVA test? 2. How can you use the F, Pvalue, and F crit values from the Excel ANOVA: Single Factor procedure to draw a conclusion? 3. What assumptions does ANOVA rely upon?
Chapter 7 Statistical Inference
297
ChiSquare Test for Independence A common problem in business is to determine whether two categorical variables are independent. We introduced the concept of independent events in Chapter 5. In the energy drink survey example (Example 5.11), we used conditional probabilities to determine whether brand preference was independent of gender. However, with sample data, sampling error can make it difficult to properly assess the independence of categorical variables. We would never expect the joint probabilities to be exactly the same as the product of the marginal probabilities because of sampling error even if the two variables are statistically independent. Testing for independence is important in marketing applications.
EXAMPL E 7.15
Independence and Marketing Strategy
Figure 7.16 shows a portion of the sample data used in Chapter 5 for brand preferences of energy drinks (Excel file Energy Drink Survey) and the crosstabulation of the results. A key marketing question is whether the proportion of males who prefer a particular brand is no different from the proportion of females. For instance, of the 63 males, 25 (40%) prefer brand 1. If gender and brand preference are indeed independent, we would expect that about the same proportion of the sample of females would also prefer brand 1. In actuality, only 9 of 37 (24%) prefer brand 1. Similarly,
27% of males prefer brand 2 versus 16% of females; and 33% of males prefer brand 3 versus 59% of females. However, we do not know whether these differences are simply due to sampling error or represent a significant difference. Knowing whether gender and brand preference are independent can help marketing personnel better target advertising campaigns. If they are not independent, then advertising should be targeted differently to males and females, whereas if they are independent, it would not matter.
We can test for independence by using a hypothesis test called the chisquare test for independence. The chisquare test for independence tests the following hypotheses: H0: the two categorical variables are independent H1: the two categorical variables are dependent The chisquare test is an example of a nonparametric test, that is, one that does not depend on restrictive statistical assumptions, as ANOVA does. This makes it a widely applicable and popular tool for understanding relationships among categorical data. The first step in the procedure is to compute the expected frequency in each cell of
▲▲ Figure
7.16
Portion of Energy Drink Survey and CrossTabulation
298
Chapter 7 Statistical Inference
▶▶ Figure
7.17
Expected Frequencies for the ChiSquare Test
the crosstabulation if the two variables are independent. This is easily done using the following: Expected Frequency in Row i and Column j =
EXAMPLE 7.16
(Grand Total Row i)(Grand Total Column j) Total Number of Observations (7.7)
Computing Expected Frequencies
For the Energy Drink Survey data, we may compute the expected frequencies using the data from the crosstabulation and formula (7.7). For example, the expected frequency of females who prefer brand 1 is (37) (34) , 100 ∙ 12.58. This
can easily be implemented in Excel. Figure 7.17 shows the results (see the Excel file ChiSquare Test). The formula in cell F11, for example, is = $I5*F$7/$I$7, which can be copied to the other cells to complete the calculations.
Next, we compute a test statistic, called a chisquare statistic, which is the sum of the squares of the differences between observed frequency, fo, and expected frequency, fe, divided by the expected frequency in each cell:
x2 = a
1fo  fe 2 2 (7.8) fe
The closer the observed frequencies are to the expected frequencies, the smaller the value of the chisquare statistic will be. The sampling distribution of x2 is a special distribution called the chisquare 1x2 2 distribution. The chisquare distribution is characterized by degrees of freedom, similar to the tdistribution. Table A.3 in Appendix A in the back of this book provides critical values of the chisquare distribution for selected values of a. We compare the chisquare statistic for a specified level of significance a to the critical value from a chisquare distribution with 1r  121c  12 degrees of freedom, where r and c are the number of rows and columns in the crosstabulation table, respectively. The Excel function CHISQ.INV.RT(probability, deg_freedom) returns the value of x2 that has a righttail area equal to probability for a specified degree of freedom. By setting probability equal to the level of significance, we can obtain the critical value for the hypothesis test. If the test statistic exceeds the critical value for a specified level of significance, we reject H0. The Excel function CHISQ.TEST(actual_range, expected_range) computes the pvalue for the chisquare test.
Chapter 7 Statistical Inference
▶▶ Figure
299
7.18
Excel Implementation of ChiSquare Test
EXAMPL E 7.17
Conducting the ChiSquare Test for Independence
For the Energy Drink Survey data, Figure 7.18 shows the calculations of the chisquare statistic using formula (7.8). For example, the formula in cell F17 is = (F5  F11)^2>F11, which can be copied to the other cells. The grand total in the 2 lower right cell is the value of x . In this case, the chisquare test statistic is 6.4924. Since the crosstabulation has r = 2 rows and c = 3 columns, we have (2  1) (3  1) = 2 degrees of freedom for the chisquare distribution. Using a = 0.05, the Excel function CHISQ.INV.RT(0.05, 2) returns the critical value 5.99146. Because the test statistic exceeds the critical value, we reject the null hypothesis that the two categorical variables are independent.
Alternatively, we could simply use the CHISQ.TEST function to find the pvalue for the test and base our conclusion on that without computing the chisquare statistic. For this example, the function CHISQ.TEST(F5:H6, F11:H12) returns the pvalue of 0.0389, which is less than a = 0.05; therefore, we reject the null hypothesis that brand and gender are independent. Because a higher proportion of females prefer brand 3 while a higher proportion of males prefer brand 1, marketers might advertise brand 3 in femalerelated media, while brand 1 should receive higher emphasis in malerelated media.
Cautions in Using the ChiSquare Test First, when using PivotTables to construct a crosstabulation and implement the chisquare test in Excel similar to Figure 7.17, be extremely cautious of blank cells in the PivotTable. Blank cells will not be counted in the chisquare calculations and will lead to errors. If you have blank cells in the PivotTable, simply replace them with zeros, or right click in the PivotTable, choose PivotTable Options, and enter 0 in the field for the checkbox For empty cells show. Second, the chisquare test assumes adequate expected cell frequencies. A rule of thumb is that there be no more than 20% of cells with expected frequencies smaller than 5, and no expected frequencies of zero. More advanced statistical procedures exist to handle this, but you might consider aggregating some of the rows or columns in a logical fashion
300
Chapter 7 Statistical Inference
to enforce this assumption, as the Example 7.18 illustrates. This, of course, results in fewer rows or columns, which changes the number of degrees of freedom. EXAMPLE 7.18
Violations of ChiSquare Assumptions
A survey of 100 students at a university queried their beverage preferences at a local coffee shop. The results are shown in the table below. Brewed coffee Female Male Total
Iced coffee
Espresso 10 5 15
6 16 22
Cappuccino
Latte
2 8 10
Mocha
4 2 6
7 3 10
9 2 11
Iced blended Tea 10 7 17
8 1 9
Total
6.16 4.84 11
Iced blended Tea 9.52 7.48 17
5.04 3.96 9
56 44 100
The expected frequencies are shown next. Expected frequencies Brewed coffee Iced coffee Female 12.32 Male 9.68 Total 22
Espresso 8.4 6.6 15
Cappuccino 5.6 4.4 10
Latte 3.36 2.64 6
Mocha 5.6 4.4 10
Total 56 44 100
If we were to conduct a chisquare test of independence, we would see that of the 16 cells, five, or over 30%, have frequencies smaller than 5. Four of them are in the Cappuccino, Latte, and Mocha columns; these can be aggregated into one column called Hot Specialty beverages. Brewed coffee Female Male
Iced coffee
Espresso 10 5 15
6 16 22
Hot Specialty 2 8 10
Iced blended 20 7 27
Tea
Total
10 7 17
8 1 9
56 44 100
Now only 2 of 12 cells have an expected frequency less than 5; this now meets the assumptions of the chisquare test. Expected frequencies Brewed coffee Female Male
Iced coffee 12.32 9.68 22
Espresso 8.4 6.6 15
Hot Specialty 5.6 4.4 10
15.12 11.88 27
Iced blended Tea 9.52 7.48 17
Total 5.04 3.96 9
56 44 100
ChiSquare Goodness of Fit Test In Chapter 5, we introduced the chisquare goodness of fit test to determine whether data can be reasonably assumed to come from a specified distribution like the normal or exponential distribution. What we actually did was to conduct a hypothesis test: H0: data are sampled from a specified distribution H1: data are not sampled from the specified distribution You can see the similarity to the chisquare test for independence in how the actual and expected frequencies are used to calculate the chisquare statistic. If the chisquare statistic is larger than the critical value, reject the null hypothesis; otherwise, fail to reject it and conclude that the specified distribution is a reasonable model for the sample data. The key difference is how we used the hypothesized probability distribution to compute the expected frequencies.
Chapter 7 Statistical Inference
301
CHECK YOUR UNDERSTANDING 1. What hypothesis does the chisquare test for independence test? 2. Why is it important to understand independence of categorical variables in business applications? 3. Explain the steps required to conduct chisquare tests for independence and normality in Excel. 4. What cautions should you be aware of in using the chisquare test for independence?
ANALYTICS IN PRACTICE: Using Hypothesis Tests and Business Analytics in Schlumberger is an international oilfieldservices provider headquartered in Houston, Texas. Through an outsourcing contract, they supply help desk services for a global telecom company that offers wireline communications and integrated telecom services to more than 2 million cellular subscribers. The help desk, located in Ecuador, faced increasing customer complaints and losses in dollars and cycle times. The company drew upon the analytics capability of one of the help desk managers to investigate and solve the problem. The data showed that the average solution time for issues reported to the help desk was 9.75 hours. The company set a goal to reduce the average solution time by 50%. In addition, the number of issues reported to the help desk had reached an average of 30,000 per month. Reducing the total number of issues reported to the help desk would allow the company to address those issues that hadn’t been resolved because of a lack of time, and to reduce the number of abandoned calls. They set a goal to identify preventable issues so that customers would not have to contact the help desk in the first place, and set a target of 15,000 issues. As part of their analysis, they observed that the average solution time for help desk technicians working at the call center seemed to be lower than the average for technicians working on site with clients. They conducted a hypothesis test structured around the question: Is there a difference between having help desk employees working at an offsite facility rather than on site within the client’s main office? The null hypothesis was that there was no significant difference; the alternative hypothesis was
Hurst Photo/Shutterstock
a Help Desk Service Improvement Project2
that there was a significant difference. Using a twosample ttest to assess whether the call center and the help desk are statistically different from each other, they found no statistically significant advantage in keeping help desk employees working at the call center. As a result, they moved help desk agents to the client’s main office area. Using a variety of other analytical techniques, they were able to make changes to their process, resulting in the following: ■■ ■■
■■
■■
■■
a decrease in the number of help desk issues of 32% improved capability to meet the target of 15,000 total issues a reduction in the average desktop solution time from 9.75 hours to 1 hour, an improvement of 89.7% a reduction in the callabandonment rate from 44% to 26% a reduction of 69% in help desk operating costs
2 Based on Francisco, Endara M. “Help Desk Improves Service and Saves Money with Six Sigma,” American Society for Quality, http:// asq.org/economiccase/markets/pdf/helpdesk24490.pdf, accessed 8/19/11.
302
Chapter 7 Statistical Inference
KEY TERMS Alternative hypothesis Analysis of variance (ANOVA) Chisquare distribution Chisquare statistic Confidence coefficient Factor Hypothesis Hypothesis testing Level of significance
Null hypothesis Onesample hypothesis test Onetailed test of hypothesis pValue (observed significance level) Power of the test Statistical inference Twotailed test of hypothesis Type I error Type II error
CHAPTER 7 TECHNOLOGY HELP Useful Excel Functions and Formulas NORM.S.INV(1  a) Finds the critical value for a onesample, onetailed test when the standard deviation is known. For a lowertailed test, the critical value is negative; for a twotailed test, use a> 2. T.INV(1  a, n  1) Finds the critical value for a onesample, onetailed test for the mean when the population standard deviation is unknown. For a lowertailed test, the critical value is negative; for a twotailed test, use a> 2. T.INV.2T(a, n  1) Finds the critical value for a onesample, twotailed test of hypothesis when the population standard deviation is unknown. The critical value for the lower tail is negative. Note that this function is the same as T.INV(1  a / 2, n  1). T.DIST(t, n  1, TRUE) Finds the pvalue for a lowertailed hypothesis test for the mean when the population standard deviation is unknown. 1  T.DIST(t, n  1, TRUE) Finds the pvalue for an uppertailed hypothesis test for the mean when the population standard deviation is unknown using the test statistic t. T.DIST.2T(ABS(t), n  1) Finds the pvalue for a twotailed hypothesis test for the mean when the population standard deviation is unknown for the test statistic t. NORM.S.DIST(z, TRUE) Finds the pvalue for a lowertailed test for a proportion using the test statistic z. 1  NORM.S.DIST(z, TRUE) Finds the pvalue for an uppertailed test for a proportion using the test statistic z.
2*NORM.S.DIST(ABS(z), TRUE) Finds the pvalue for a twotailed test for a proportion using the test statistic z. CHISQ.INV.RT(probability, deg_ freedom) Returns the value of chisquare that has a righttail area equal to probability for a specified degree of freedom. By setting probability equal to the level of significance, we obtain the critical value for the chisquare hypothesis test for independence. CHISQ.TEST(actual_range, expected_range) Computes the pvalue for the chisquare test for independence. Excel Templates OneSample Hypothesis Tests (Figure 7.6): Open the Excel file One Sample Hypothesis Tests. Select the worksheet tab corresponding to the type of hypothesis test to perform and enter the appropriate data. Excel Techniques Excel zTest: TwoSample for Means: Performs a twosample hypothesis test for means when s2 for each population is known. Select zTest: TwoSample for Means from the Data Analysis menu and enter the data ranges, hypothesized mean difference, and value of a. Check Labels if headers are included. Excel tTest: TwoSample Assuming Unequal Variances (Example 7.10): Performs a twosample test for means when s2 for each population is unknown and assumed unequal. Select tTest: TwoSample Assuming Unequal Variances from the Data Analysis menu and enter
Chapter 7 Statistical Inference
the data ranges, hypothesized mean difference, and value of a. Check Labels if headers are included. Excel tTest: TwoSample Assuming Equal Variances: Performs a twosample test for means when s2 for each population is unknown and assumed equal. Select tTest: TwoSample Assuming Equal Variances from the Data Analysis menu and enter the data ranges, hypothesized mean difference, and value of a. Check Labels if headers are included. For each of the above tests, apply the following rules: 1. If the test statistic is negative, the onetailed pvalue is the correct pvalue for a lowertailed test; however, for an uppertailed test, you must subtract this number from 1.0 to get the correct pvalue. 2. If the test statistic is nonnegative (positive or zero), then the pvalue in the output is the correct pvalue for an uppertailed test, but for a lowertailed test, you must subtract this number from 1.0 to get the correct pvalue. 3. For a lowertailed test, you must change the sign of the onetailed critical value. Excel tTest: Paired TwoSample for Means (Example 7.11): Performs a twosample test for means using paired samples. Select tTest: Paired TwoSample for Means from the Data Analysis menu and enter the data ranges, hypothesized mean difference, and value of a. Check Labels if headers are included. Excel FTest: TwoSample for Variances (Example 7.12): Performs a twosample test for equality of v ariances. Select FTest: TwoSample for Variances from the Data Analysis menu and enter the data ranges, hypothesized mean difference, and value of a. Make sure that the first variable has the larger variance and enter a> 2, not a, for the level of significance in the dialog. Check Labels if headers are included. ANOVA (Example 7.14): Set up the worksheet so that the data you wish to use are displayed in contiguous columns. Select ANOVA: Single Factor from the Data Analysis menu. Enter the input range as the rectangular array that contains all the data; check Labels in the First Row if headers are included. StatCrunch StatCrunch provides the ability to conduct the hypothesis tests, analysis of variance, and chisquare tests that we
303
learned about in this chapter. You can find video tutorials and stepbystep procedures with examples at https:// www.statcrunch.com/5.0/example.php. We suggest that you first view the tutorials Getting started with StatCrunch and Working with StatCrunch sessions. The following tutorials explain how to conduct hypothesis tests in StatCrunch. These are located under the headings Traditional Hypothesis Tests For, ANOVA, and Summary Statistics and Tables. ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■ ■■
A mean with raw data A mean with summary data A proportion with raw data A proportion with summary data The difference between two means with raw data The difference between two means with summary data The difference between means with paired data The difference between two proportions with raw data The difference between two proportions with summary data One way ANOVA Contingency tables from raw data Contingency tables from summary data
Example: Testing a Hypothesis for a Mean with Raw Data 1. Choose the Stat 7 T Stats 7 One Sample 7 With Data menu option. 2. Select the column containing the sample data values. 3. Under Perform, select Hypothesis test for m . 4. Enter the null mean and choose ∙, 6 , or 7 for the alternative. 5. Click Compute! to view the results. Example: Conducting a TwoSample Hypothesis Test for the Mean with Summary Data 1. Choose the Stat 7 T Stats 7 One Sample 7 With Data menu option. 2. Choose the With Summary option to enter the sample mean, sample standard deviation, and sample size for both samples. 3. Uncheck the Pool variances option if desired. 4. Select the Hypothesis test option, enter the difference in means for the null hypothesis, and choose ∙, 6 , or 7 for the alternative. 5. Click Compute! to view the results.
304
Chapter 7 Statistical Inference
PROBLEMS AND EXERCISES For all hypothesis tests, assume that the level of significance is 0.05 unless otherwise stated. Hypothesis Testing 1. When President Donald Trump took office, he
believed that the reason he did not win the popular vote was because 3 to 5 million people voted illegally. Explain how hypothesis testing might be used in a similar fashion as the legal analogy example. OneSample Hypothesis Tests 2. A gym trainer in Cologne, Germany, believes that the
average time spent exercising each day by the city is greater than 1.5 hours. Formulate an appropriate onesample test of hypothesis to test this belief. 3. A research group in Sweden estimated that almost
1.5 million people in Sweden have used Netflix from September to December 2018. Formulate an appropriate onesample test of hypothesis to test this. 4. Looking at the data and information in the Excel file
Car Sharing Survey, state some examples of interesting hypothesis tests by proposing null and alternative hypotheses similar to those in Example 7.2. 5. Looking at the data and information in the Excel
file Retail Survey, state some examples of interesting hypothesis tests by proposing null and alternative hypotheses similar to those in Example 7.2. 6. Looking at the data and information in the Excel file
TV Viewing Survey, state some examples of interesting hypothesis tests by proposing null and alternative hypotheses similar to those in Example 7.2. 7. A pharmaceutical company in Australia is testing a
medical drug that intends to increase heart rate. A sample of 124 tests yielded a mean increase of 1.4 beats per minute, with a population standard deviation known to be 3.6. Since the company wants to avoid marketing an ineffective medical drug, it proposes a 0.001 significance level. If the drug does not work, increase will be zero. Formulate and conduct a hypothesis test to determine if you can conclude that the company should market the drug. 8. An analyst at a business school in France is assess
ing students’ performance. Historical records of applicants suggest that the mean score obtained by students attending the school is 350. The business
school administration wants to know whether the performance of students enrolled in the current year has changed or not. From the recent scores of a sample of 100 students, the average score is 365 with a standard deviation of 38. Does this data provide statistical evidence that the quality of students has improved? Formulate the appropriate hypothesis test and conduct the test. 9. According to a magazine, people read an average of
more than three books in a month. A survey of 20 random individuals found that the mean number of books they read was 2.9 with a standard deviation of 1.23. a. To test the magazine’s claim, what should the appropriate hypotheses be? b. Compute the test statistic. c. Using a level of significance of 0.05, what is the critical value? d. Find the pvalue for the test. e. What is your conclusion? 10. A marine biologist wants to use a parrotfish for research
purposes as she believes that the variability of weights among parrotfish is small. From a random sample of 50 parrotfish, she found that their mean weight is 6.3 pounds with a standard deviation of 2.4 pounds. She will only use the parrotfish if there is evidence that the variance of their weights is less than 4 pounds. Formulate an appropriate hypothesis test to determine if there is enough evidence that the weight is small enough for the marine biologist to use them in her research. 11. A pharmacist believes that in manufacturing a pill or
tablet medicine, the need for an active ingredient in each pill is less than 14 milligrams. She selects a random sample of 15 pills and determines that each pill has an average of 10.5 milligrams of the active ingredient with a standard deviation of 0.56 milligrams. Formulate the appropriate hypothesis test and draw a conclusion. 12. Using the data in the Excel file Consumer Transpor
tation Survey, test the following null hypotheses: a. Individuals spend at least eight hours per week in their vehicles. b. Individuals drive an average of 600 miles per week. c. The average age of SUV drivers is no greater
than 35.
Chapter 7 Statistical Inference
d. At least 80% of individuals are satisfied with
their vehicles. Perform the calculations using the correct formulas and Excel functions, and compare your results with the One Sample Hypothesis Test Excel template to verify them. 13. A gum manufacturer claims that on average the fla
vor of an entire packet of its gum would last for more than 39 minutes. A quality controller selects a random sample of 55 packets of gum. She finds the average time for which the gum flavor lasts is 40 minutes with a standard deviation of 5.67 minutes. a. Formulate a hypothesis test to validate the manufacturer’s claim. b. After a new technique to improve the lasting
period of the gum flavor was applied, the quality controller reselected 60 packets of gum and found that the average time for which the gumflavor lasts is 45 minutes with a standard deviation of 3.15 minutes. Is there sufficient evidence to conclude that the new technique significantly increased the lasting time? c. Use a 95% confidence interval for the population average time for which the flavor lasts to validate the manufacturer’s claim after the new technique is applied. 14. A diet program claims that the average amount of water
that a person on a diet should drink each day to lose weight is 4 liters. A random sample of 13 people enrolled in the diet program found that the average consumption is 3.6 liters of water per day with a standard deviation of 0.379 liters. Is the diet program’s claim valid? 15. Call centers typically have high turnover. The director
of human resources for a large bank has compiled data on about 70 former employees at one of the bank’s call centers in the Excel file Call Center Data 2. In writing an article about the working conditions at call centers, a reporter claimed that the average tenure is no more than two years. Formulate and test a hypothesis using these data to determine if this claim can be disputed. 16. Using the data in the Excel file Airport Service Times,
determine if the airline can claim that its average service time is less than 2.5 minutes. 17. A computer repair firm believes that its average
repair time is less than two weeks. Using the data in the Excel file Computer Fix Times, determine if the company can continue to support this claim.
305
18. A survey on female health stated that the standard
body fat percentage of a woman of the average age of 25 years with an average weight of 140 pounds and an average height of 64 inches is 16%. The Excel file Women Body Fat Percentage provides the data for this survey. Use the dataset to conduct a test of hypothesis for each of the variables captured. 19. Formulate and test hypotheses to determine if sta
tistical evidence suggests that the graduation rate for (1) top liberal arts colleges or (2) research universities in the Excel file Colleges and Universities exceeds 90%. Do the data support a conclusion that the graduation rates exceed 85%? Would your conclusions change if the level of significance was 0.01 instead of 0.05? 20. An industry trade publication stated that the average
profit per customer for this industry was greater than $4,500. The Excel file Sales Data provides data on a sample of customers. Using a test of hypothesis, do the data support this claim or not? 21. A financial advisor believes that the proportion of
investors who are riskaverse (that is, try to avoid risk in their investment decisions) is at least 0.6. A survey of 31 investors found that 19 of them were riskaverse. Formulate a onesample hypothesis test for a proportion to test this belief. 22. Suppose that a previous record of a blood donation
campaign conducted in the city of London indicates that the probability of a donor having blood type O is 0.6. Of a sample of 120 donors who have come to donate blood at this year’s blood donation camp, 70 donors have blood type O. Assume that the blood donors are not related in any way, so that we can consider them independent. Can the management of the blood donation campaign conclude that they have had the same result as the previous record for the proportion of donors having blood type O? 23. The Excel file Room Inspection provides data for 100
room inspections at each of 25 hotels in a major chain. Management would like the proportion of nonconforming rooms to be less than 2%. Formulate a onesample hypothesis test for a proportion and perform the calculations using the correct formulas and Excel functions, and compare your results with the One Sample Hypothesis Test Excel template to verify them. 24. An online bookseller is considering selling an
ereader but will do so only if they have evidence that the proportion of customers who will likely purchase
306
Chapter 7 Statistical Inference
one is more than 0.4. Of a survey of 25 customers, 8 of them stated that they would likely purchase an ereader. What should the bookseller do? 25. An employer is considering negotiating its pricing
structure for health insurance with its provider if there is sufficient evidence that customers will be willing to pay a lower premium for a higher deductible. Specifically, they want at least 30% of their employees to be willing to do this. Using the sample data in the Excel file Insurance Survey, determine what decision they should make. TwoSample Hypothesis Tests 26. A twosample test for means was conducted to test
whether the mean number of movies watched each month differed between males and females. The Excel Data Analysis tool results are shown below. tTest: TwoSample Assuming Unequal Variances Female
Male
Mean
5.6
7.5
Variance
6.267
21.833
Observations
10
10
Hypothesized Mean Difference 0 df
14
t Stat
1.133
P(T 6=t) onetail
0.138
t Critical onetail
1.76
P(T 6=t) twotail
0.276
t Critical twotail
2.144
a. Explain how to use this information to
draw a conclusion if the null hypothesis is H0: mF  mM … 0. Clearly state the correct critical value and pvalue and your conclusion. b. Explain how to use this information to draw a conclusion if the null hypothesis is H0: mF  mM Ú 0. Clearly state the correct critical value and pvalue and your conclusion. c. Explain how to use this information to draw a conclusion if the null hypothesis is H0: mF  mM = 0. Clearly state the correct critical value and pvalue and your conclusion. 27. A twosample test for means was conducted to deter
mine if the completion time for continuing education programs for nurses differed when costs are paid by employers (Yes) versus when individuals paid out
of his or her own funds (No). The following are the Excel Data Analysis tool results. tTest: TwoSample Assuming Unequal Variances Mean
Yes
No
33.982
28.556
Variance
448.127
349.504
Observations
12
7
Hypothesized Mean Difference
0
df
14
t Stat
0.580
P(T 6=t) onetail
0.286
t Critical onetail
1.761
P(T 6=t) twotail
0.571
t Critical twotail
2.145
a. Explain how to use this information to
draw a conclusion if the null hypothesis is H0: mY  mN … 0. Clearly state the correct critical value and pvalue and your conclusion. b. Explain how to use this information to draw a conclusion if the null hypothesis is H0: mY  mN Ú 0. Clearly state the correct critical value and pvalue and your conclusion. c. Explain how to use this information to draw a conclusion if the null hypothesis is H0: mY  mN = 0. Clearly state the correct critical value and pvalue and your conclusion. 28. The director of human resources for a large bank has
compiled data on about 70 former employees at one of the bank’s call centers (see the Excel file Call Center Data 2). For each of the following, assume equal variances of the two populations. a. Test the null hypothesis that the average length of service for males is the same as for females. b. Test the null hypothesis that the average length of service for individuals without prior call center experience is the same as those with experience. c. Test the null hypothesis that the average length of service for individuals with a college degree is the same as for individuals without a college degree. d. Now conduct tests of hypotheses for equality of variances. Were your assumptions of equal variances valid? If not, repeat the test(s) for means using the unequal variance test.
Chapter 7 Statistical Inference
29. Using the Excel file YouTube Hours, determine if the
mean number of hours spent online per week is the same for males as it is for females. 30. Determine if there is evidence to conclude that the
mean number of vacations taken by married individuals is less than the number taken by single/divorced individuals using the data in the Excel file Vacation Survey. Use a level of significance of 0.05. Would your conclusion change if the level of significance were 0.01? 31. The Excel file Fortune100 Accountants provides the
307
Q2: Did the support representative keep you informed about the status of progress in resolving your problem? Q3: Was the support representative courteous and professional? Q4: Was your problem resolved? Q5: Was your problem resolved in an acceptable amount of time? Q6: Overall, how did you find the service provided by our technical support department?
results of a survey of 35 employees in a tax division of a Fortune 100 company. a. Is there sufficient evidence at the 0.01 level of significance that the average number of years of service is the same for both genders?
A final question asks the customer to rate the overall quality of the product using a scale of 0—very poor; 1—poor; 2—good; 3—very good; 4—excellent. A sample of survey responses and associated resolution and response data are provided in the Excel file Customer Support Survey.
b. Is there sufficient evidence at the 0.01 level of
a. The company has set a service standard of one
significance that the average years of undergraduate study is the same for both genders? ences, conduct a hypothesis test to determine if price and taste ratings are the same for large/chain stores versus small/independent coffee shops.
day for the mean resolution time. Does evidence exist that the resolution time is more than one day? How do the outliers in the data affect your result? What should you do about them? b. Test the hypothesis that the average service index is equal to the average engineer index.
33. In the Excel file Cell Phone Survey, test the hypoth
38. Using the data in the Excel file Ohio Education Per
esis that the mean responses for value for the dollar and customer service do not differ by gender.
formance, test the hypotheses that the mean difference in writing and reading scores is zero and that the mean difference in math and science scores is zero. Use the pairedsample procedure.
32. For the data in the Excel file Coffee Shop Prefer
34. In the Excel file Credit Risk Data, test the hypotheses that
the number of months employed is the same for applicants with low credit risk as for those with high credit risk evaluations. Use a level of significance of 0.01.
39. The Excel file Unions and Labor Law Data reports
fall in area A is significantly more than that of area B using the data in the Excel file Rainfall.
the percent of public and privatesector employees in unions in 1982 for each state, along with indicators of whether the states had a bargaining law that covered public employees or righttowork laws. a. Test the hypothesis that the mean percent of employees in unions for both the public sector and private sector is the same for states having bargaining laws as for those who do not.
37. A producer of computeraided design software for
b. Test the hypothesis that the mean percent of
the aerospace industry receives numerous calls for technical support. Tracking software is used to monitor response and resolution times. In addition, the company surveys customers who request support using the following scale: 0—did not meet expectations; 1—marginally met expectations; 2—met expectations; 3—exceeded expectations; 4—greatly exceeded expectations. The questions are as follows: Q1: Did the support representative explain the process for resolving your problem?
employees in unions for both the public sector and private sector is the same for states having righttowork laws as for those who do not.
35. For the data in the Excel file Netflix Views Survey,
conduct a hypothesis test to determine if married individuals watch less Netflix than single individuals. 36. Determine if there is evidence to conclude that rain
40. Using the data in the Excel file Student Grades,
which represent exam scores in one section of a large statistics course, test the hypothesis that the variance in grades is the same for both tests. 41. In the Excel file Restaurant Sales, determine if the
variance of weekday sales is the same as that of
308
Chapter 7 Statistical Inference
weekend sales for each of the three variables (lunch, dinner, and delivery).
d. Find the chisquare critical value and pvalue and
draw a conclusion. 48. The crosstabulation data given below represent the
Analysis of Variance (ANOVA) 42. For the Excel file Job Satisfaction, use ANOVA to
determine if the mean overall job satisfaction ratings differ by department. 43. A college is trying to determine if there is a signifi
cant difference in the mean GMAT score of students from different undergraduate backgrounds who apply to the MBA program. The Excel file GMAT Scores contains data from a sample of students. What conclusion can be reached using ANOVA?
number of passed and failed results of male and female students at a university for the last semester. a. Write the hypotheses for the chisquare test for independence. b. Find the expected frequencies. c. Compute the chisquare test statistic. d. Find the chisquare critical value and pvalue and draw a conclusion. Result
44. For the data in the Excel file Helpdesk Survey, deter
mine if the mean overall satisfaction ratings differ by the ratings given for response time. 45. Using the data in the Excel file Broadband Service
Providers, apply ANOVA to determine if the mean response for value for the pound is the same for different types of broadband service providers in the United Kingdom. 46. For the data in the Excel file Freshman College
Data, a. use ANOVA to determine whether significant differences exist in the mean retention rate for the different colleges over the fouryear period. b. use ANOVA to determine if significant differences exist in the mean ACT scores among the different colleges. c. use ANOVA to determine if significant differences exist in the mean SAT scores among the different colleges. ChiSquare Tests for Independence and Normality 47. The crosstabulation data given below represent the
number of males and females in a work group who feel overstressed and those who don’t. Overstressed
Women
Men
No
9
4
Yes
6
9
a. Write the hypotheses for the chisquare test for
independence. b. Find the expected frequencies. c. Compute the chisquare statistic using a level of significance of 0.05.
Male
Female
Pass
80
72
Fail
18
3
49. The following crosstabulation shows the number of
people who rated a customer service representative as friendly and polite based on whether the representative greeted them first. a. Write the hypotheses for the chisquare test for independence. b. Find the expected frequencies. c. Compute the chisquare statistic using a level of significance of 0.01. d. Find the chisquare critical value and pvalue and draw a conclusion. Friendly/Polite Staff Greeting
No
Yes
No
13
7
Yes
12
22
50. Conduct the chisquare test for independence using
the aggregated results for Example 7.18 using a level of significance of 0.05. 51. For the following crosstabulation data, which rep
resent a count of the type of books that individuals prefer, verify that the assumptions of the chisquare test are not met, and aggregate the data in an appropriate way so that the assumptions are met. Then conduct a chisquare test for independence. Gender Fiction Nonfiction
Graphic Novel
Autobiography
Male
11
2
2
8
Female
8
6
5
0
Chapter 7 Statistical Inference
52. A manufacturer of automobiles is bringing out a
new model. To create its advertising campaign, the manufacturer wants to determine whether the model’s appeal will be dependent on a particular age group. A sample of a customer survey revealed the following: Appeal Under 20
20–40
40–50
50 and over
Like
140
70
70
25
Dislike
60
40
30
65
Can we conclude that there is an association between the person who answers the call and customer satisfaction? 54. For the data in the Excel file Graduate School Sur
vey, perform a chisquare test for independence to determine if plans to attend graduate school are independent of gender. 55. For the data in the Excel file New Account Process
ing, perform chisquare tests for independence to determine if certification is independent of gender and if certification is independent of having prior industry background.
What can the manufacturer conclude? 53. Four call service representatives were monitored over
a period of three months. After each call, every customer was asked whether they were satisfied with the service by answering “Yes” or “No.” The following data were obtained:
56. Conduct a chisquare test to determine if a normal
distribution can reasonably model the data in the Excel file Airport Service Times. 57. Conduct a chisquare test to determine if a normal
distribution can reasonably model the expense data in the Excel file Travel Expenses.
Representative Satisfied
Dion
Susan
Alice
Anthony
Yes
172
118
150
140
No
28
82
50
16
309
58. Conduct a chisquare test to determine if a normal
distribution can reasonably model the sum of the midterm and final exam grades in the Excel file Student Grades.
CASE: DROUT ADVERTISING RESEARCH PROJECT The background for this case was introduced in Chapter 2. This is a continuation of the case in Chapter 6. For this part of the case, propose and test some meaningful hypotheses that will help Ms. Drout understand and explain the results. Include twosample tests, ANOVA, and/or chisquare tests for independence as appropriate. Write up your conclusions in a formal report, or add your findings to the
report you completed for the case in Chapter 6 as per your instructor’s requirements. If you have accumulated all sections of this case into one report, polish it up so that it is as professional as possible, drawing final conclusions about the perceptions of the role of advertising in the reinforcement of gender stereotypes and the impact of empowerment advertising.
CASE: PERFORMANCE LAWN EQUIPMENT Elizabeth Burke has identified some additional questions she would like you to answer using the Performance Lawn Equipment Database. 1. Are there significant differences in ratings of specific product/service attributes in the Customer Survey worksheet? 2. In the worksheet OnTime Delivery, has the proportion of ontime deliveries in 2018 significantly improved since 2014? 3. Have the data in the worksheet Defects After Delivery changed significantly over the past five years? 4. Although engineering has collected data on alternative process costs for building transmissions in the
worksheet Transmission Costs, why didn’t they reach a conclusion as to whether one of the proposed processes is better than the current process? 5. Are there differences in employee retention due to gender, college graduation status, or whether the employee is from the local area in the data in the worksheet Employee Retention? Conduct appropriate statistical analyses and hypothesis tests to answer these questions and summarize your results in a formal report to Ms. Burke.
This page intentionally left blank
CHAPTER
8
Trendlines and Regression Analysis
Luca Bertolli/123RF
LEARNING OBJECTIVES ■■
■■
■■
After studying this chapter, you will be able to:
Explain the purpose of regression analysis and provide examples in business. Use a scatter chart to identify the type of relationship between two variables. List the common types of mathematical functions used in predictive modeling.
■■
■■
■■
Use the Excel Trendline tool to fit models to data. Explain how leastsquares regression finds the bestfitting regression model. Use Excel functions to find leastsquares regression coefficients.
312 ■■
■■
■■
■■
■■
■■ ■■
Chapter 8 Trendlines and Regression Analysis
Use the Excel Regression tool for both single and multiple linear regressions. Interpret the regression statistics of the Excel Regression tool. Interpret significance of regression from the Excel Regression tool output. Draw conclusions for tests of hypotheses about regression coefficients. Interpret confidence intervals for regression coefficients. Calculate standard residuals. List the assumptions of regression analysis and describe methods to verify them.
■■
■■
■■
■■
■■
■■
Explain the differences in the Excel Regression tool output for simple and multiple linear regression models. Apply a systematic approach to build good regression models. Explain the importance of understanding multicollinearity in regression models. Build regression models for categorical data using dummy variables. Test for interactions in regression models with categorical variables. Identify when curvilinear regression models are more appropriate than linear models.
The late management and quality guru Dr. W. Edwards Deming once stated that all management is prediction. He was implying that when managers make decisions, they must consider what will happen in the future. For example, financial analysts cannot make good investment decisions without predicting the future movement of stock prices, and airlines cannot determine when to purchase jet fuel—and how much—without predicting how oil prices will change. Trendlnes show the movement of such attributes over time. Many other applications of predictive analytics involve modeling relationships between one or more independent variables and some dependent variable (called regression analysis). For example, we might wish to predict the level of sales based on the price we set, or to predict sales based on the U.S. GDP (gross domestic product) and the tenyear treasury bond rate to capture the influence of the business cycle.1 A marketing researcher might want to predict the intent of buying a particular automobile model based on a survey that measures consumer attitudes toward the brand, negative word of mouth, and income level.2 An insurance company may need to predict the number and amount of claims based on the demographics of its customers. Human resource managers might need to predict the need for different workforce skills in order to develop hiring and training plans. Fantasy sports hobbyists may want to predict the value of players based on various performance attributes. Thus, you will likely find numerous opportunities to apply the tools we learn in this chapter to your career and personal interests.
1 James R. Morris and John P. Daley, Introduction to Financial Models for Management and Planning (Boca Raton, FL: Chapman & Hall/CRC, 2009): 257. 2
Alvin C. Burns and Ronald F. Bush, Basic Marketing Research Using Microsoft Excel Data Analysis, 2nd ed. (Upper Saddle River, NJ: Prentice Hall, 2008): 450.
Chapter 8 Trendlines and Regression Analysis
313
Trendlines and regression analysis are tools for building predictive models. Our principal focus is to gain a basic understanding of how to use and interpret trendlines and regression models, statistical issues associated with interpreting regression analysis results, and practical issues in using trendlines and regression as tools for making and evaluating decisions.
Modeling Relationships and Trends in Data Understanding both the mathematics and the descriptive properties of different functional relationships is important in building predictive analytical models. We often begin by creating a chart of the data to understand them and choose the appropriate type of functional relationship to incorporate into an analytical model. For crosssectional data, we use a scatter chart; for timeseries data, we use a line chart. Common types of mathematical functions used in predictive analytical models include the following: ■■
■■
■■
■■
■■
Linear function: y = a + bx. Linear functions show steady increases or decreases over the range of x. This is the simplest type of function used in predictive models. It is easy to understand and, over small ranges of values, can approximate behavior rather well. Logarithmic function: y = ln1x2. Logarithmic functions are used when the rate of change in a variable increases or decreases quickly and then levels out, such as with diminishing returns to scale. Logarithmic functions are often used in marketing models where constant percentage increases in advertising, for instance, result in constant, absolute increases in sales. Polynomial function: y = ax2 + bx + c (second order—quadratic function), y = ax3 + bx2 + cx + d (third order—cubic function), and so on. A secondorder polynomial is parabolic in nature and has only one hill or valley; a thirdorder polynomial has one or two hills or valleys. Revenue models that incorporate price elasticity are often polynomial functions. Power function: y = axb. Power functions define phenomena that increase at a specific rate. Learning curves that express improving times in performing a task are often modeled with power functions having a 7 0 and b 6 0. Exponential function: y = abx. Exponential functions have the property that y rises or falls at constantly increasing rates. For example, the perceived brightness of a lightbulb grows at a decreasing rate as the wattage increases. In this case, a would be a positive number and b would be between 0 and 1. The exponential function is often defined as y = aex, where b = e, the base of natural logarithms (approximately 2.71828).
The Excel Trendline tool provides a convenient method for determining the bestfitting functional relationship among these alternatives for a set of data. First, click the chart to which you wish to add a trendline; this will display the Chart Tools menu. Select the Chart Tools Design tab, and then click Add Chart Element from the Chart Layouts group. From the Trendline submenu, you can select one of the options (Linear is the most common) or More Trendline Options . . . . If you select More Trendline Options, you will get the Format Trendline pane in the worksheet (see Figure 8.1). A simpler way of doing all this is to right click on the data series in the chart and choose Add trendline from the popup menu—try it! Select the radio button for the type of functional relationship you wish
314
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
8.1
Excel Format Trendline Pane
to fit to the data. Check the boxes for Display Equation on chart and Display Rsquared value on chart. You may then close the Format Trendline pane. Excel will display the results on the chart you have selected; you may move the equation and Rsquared value for better readability by dragging them to a different location. To clear a trendline, right click on it and select Delete. R2 (Rsquared) is a measure of the “fit” of the line to the data. The value of R2 will be between 0 and 1. The larger the value of R2, the better the fit. We will discuss this further in the context of regression analysis. Trendlines can be used to model relationships between variables and understand how the dependent variable behaves as the independent variable changes. For example, the demandprediction models that we introduced in Chapter 1 (Examples 1.7 and 1.8) would generally be developed by analyzing data.
EXAMPLE 8.1
Modeling a PriceDemand Function
A market research study has collected data on sales volumes for different levels of pricing of a particular product. The data and a scatter diagram are shown in Figure 8.2 (Excel file PriceSales Data). The relationship between price and sales clearly appears to be linear, so a linear trendline was fit to the data. The resulting model is
Sales = 20,512  9.5116 * Price If the price is $125, we can estimate the the level of sales as Sales = 20,512  9.5116 * 125 = 19,323 This model can be used as the demand function in other marketing or financial analyses.
Trendlines are also used extensively in modeling trends over time—that is, when the variable x in the functional relationships represents time. For example, an analyst for an airline needs to predict where fuel prices are going, and an investment analyst would want to predict the price of stocks or key economic indicators.
Chapter 8 Trendlines and Regression Analysis
EXAMPL E 8.2
Predicting Crude Oil Prices
Figure 8.3 shows a chart of historical data on crude oil prices on the first Friday of each month from January 2006 through June 2008 (data are in the Excel file Crude Oil Prices). Using the Trendline tool, we can try to fit the various functions to these data (here x represents the number of months starting with January 2006). The results are as follows: Exponential:
0.021x
y = 50.49e
2
R = 0.664
Logarithmetic: y = 13.02ln(x) + 39.60 R 2 = 0.382
▶▶ Figure
8.2
Scatter Chart with Fitted Linear Function
▶▶ Figure
315
8.3
Chart of Crude Oil Prices Data
Polynomial (second order): y = 0.130x 2  2.40x + 68.01 Polynomial (third order): y = 0.005x 3  0.111x 2 + 0.648x + 59.497 Power: y = 45.96x 0.0169
R2 = 0.905
R2 = 0.928 R 2 = 0.397
The bestfitting model among these, which has the largest R2, is the thirdorder polynomial, shown in Figure 8.4.
316
Chapter 8 Trendlines and Regression Analysis
Be cautious when using polynomial functions. The R2 value will continue to increase as the order of the polynomial increases; that is, a thirdorder polynomial will provide a better fit than a secondorder polynomial, and so on. Higherorder polynomials will generally not be very smooth and will be difficult to interpret visually. Thus, we don’t recommend going beyond a thirdorder polynomial when fitting data. Use your eye to make a good judgment! Of course, the proper model to use depends on the scope of the data. As the chart in Figure 8.3 shows, crude oil prices were relatively stable until early 2007 and then began to increase rapidly. If the early data are included, the longterm functional relationship might not adequately express the shortterm trend. For example, fitting a model to only the data beginning with January 2007 yields these models: Exponential: Polynomial (second order): Linear:
y = 50.56 e0.044x y = 0.121x2 + 1.23x + 53.48 y = 3.55x + 45.76
R2 = 0.969 R2 = 0.968 R2 = 0.944
The difference in prediction can be significant. For example, predicting the price six months after the last data point 1x = 362 yields $172.25 for the thirdorder polynomial fit with all the data and $246.45 for the exponential model with only the recent data. Thus, you must be careful to select the proper amount of data for the analysis. The question then becomes one of choosing the best assumptions for the model. Is it reasonable to assume that prices would increase exponentially, or perhaps at a slower rate, such as with the linear model fit? Or would they level off and start falling? Clearly, factors other than historical trends would enter into this choice. In the latter half of 2008, oil prices plunged; thus, all predictive models are risky. CHECK YOUR UNDERSTANDING 1. State the common types of mathematical functions used in predictive analytics and their properties. 2. Explain how to use the Trendline tool in Excel. 3. What does R2 measure?
▶▶ Figure
8.4
Polynomial Fit of Crude Oil Prices Data
Chapter 8 Trendlines and Regression Analysis
317
ANALYTICS IN PRACTICE: Using Predictive Trendline Models at
Procter & Gamble3
Procter & Gamble (P&G) laundry products are global household brands that include Tide, Dash, and Gain and are offered in several physical product forms, including powders, liquids, pods, tablets, and bars. These products are manufactured at more than 30 sites and sold in more than 150 countries worldwide. The design of laundry product formulations (that is, ingredient composition of chemical mixtures) has become more complex over the years because of challenges such as productportfolio expansion, rapidly changing ingredient costs and availability, and increasing competitive activity. P&G’s research and development organization is at the forefront of the development and adoption of modeling tools that enable the company to make better decisions on product formulation, processing, and manufacturing. These include empirical models that
predict chemical reactions during manufacturing, inuse physical properties of the product, technical performance of the product, and even consumer acceptance rates. These tools enable researchers to instantly predict a product’s physical properties and performance, integrate models, and balance production tradeoffs using a variety of predictive and prescriptive capabilities. Predictive models the company has used include thirdorder polynomial functions that capture the two performance qualities of a mixture: stain removal and whiteness. For example, stain removal performance is predicted by the stain removal index (SRI) response function, which has the following form: SRI = C0 + C1v1 + C2v2 + C3v1v2 + c, where Ci represent coefficients and vi represent design variables: such as wash concentrations (milligrams per liter) of chemical ingredients and wash conditions (for example, temperature).
Simple Linear Regression Regression analysis is a tool for building mathematical and statistical models that characterize relationships between a dependent variable (which must be a ratio variable and not categorical) and one or more independent, or explanatory, variables, all of which are numerical (but may be either ratio or categorical). Two broad categories of regression models are often used in business settings: (1) regression models of crosssectional data and (2) regression models of timeseries data, in which the independent variables are time or some function of time and the focus is on predicting the future. Timeseries regression is an important tool in forecasting, which is the subject of Chapter 9. A regression model that involves a single independent variable is called simple regression. A regression model that involves two or more independent variables is called multiple regression. In the remainder of this chapter, we describe how to develop and analyze both simple and multiple regression models. Simple linear regression involves finding a linear relationship between one independent variable, X, and one dependent variable, Y. The relationship between two variables can assume many forms, as illustrated in Figure 8.5. The relationship may be linear or nonlinear, or there may be no relationship at all. Because we are focusing our discussion on linear regression models, the first thing to do is to verify that the relationship is linear, as in F igure 8.5(a). We would not expect to see the data line up perfectly along a straight line; we simply want to verify that the general relationship is linear. If the relationship is clearly nonlinear, as in Figure 8.5(b), then alternative approaches must be used, and if no relationship is evident, as in Figure 8.5(c), then it is pointless to even consider developing a linear regression model. 3
Adapted from Nats Esquejo, Kevin Miller, Kevin Norwood, Ivan Oliveira, Rob Pratt, and Ming Zhao, “Statistical and Optimization Techniques for Laundry Portfolio Optimization at Procter & Gamble,” Interfaces, Vol. 45, No. 5, September–October 2015, pp. 444–461.
318
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
8.5
Examples of Variable Relationships
(a) Linear
(b) Nonlinear
(c) No relationship
To determine if a linear relationship exists between the variables, we recommend that you create a scatter chart that can show the relationship between variables visually. EXAMPLE 8.3
Home Market Value Data
The market value of a house is typically related to its size. In the Excel file Home Market Value (see Figure 8.6), data obtained from a county auditor provide information about the age, square footage, and current market value of houses in a particular subdivision. We might wish to investigate the relationship between the market value and the size of the home. The independent variable, X, is the number of square feet, and the dependent variable, Y, is the market value.
▶▶ Figure
8.6
Portion of Home Market Value
▶▶ Figure
8.7
Scatter Chart of Market Value Versus Home Size
Figure 8.7 shows a scatter chart of the market value in relation to the size of the home. In general, we see that higher market values are associated with larger house sizes, and the relationship is approximately linear. Therefore, we could conclude that simple linear regression would be an appropriate technique for predicting market value based on house size.
Chapter 8 Trendlines and Regression Analysis
319
Finding the BestFitting Regression Line The idea behind simple linear regression is to express the relationship between the dependent variable Y and the independent variable X by a simple linear equation, Y = a + bX where a is the yintercept and b is the slope of the line. For the home market value example, this would be Market Value = a + b * Square Feet If we draw a straight line through the data, some of the points will fall above the line, some will fall below it, and a few might fall on the line itself. Figure 8.8 shows two possible straight lines that pass through the data. Clearly, you would choose A as the betterfitting line over B because all the points are closer to the line and the line appears to be in the middle of the data. The only difference between the lines is the value of the slope and intercept; thus, we seek to determine the values of the slope and intercept that provide the bestfitting line. We can find the bestfitting line using the Excel Trendline tool (with the Linear option chosen), as described earlier in this chapter.
EXAMPL E 8.4
Using Excel to Find the Best Regression Line
When using the Trendline tool for simple linear regression in the Home Market Value example, be sure the Linear option is selected (it is the default option when you use the tool). Figure 8.9 shows the bestfitting regression line. The equation is Market Value = $32,673 + $35.036 * Square Feet The value of the regression line can be explained as follows. Suppose we wanted to estimate the home market value for any home in the population from which the sample data were gathered. If all we knew were the market values, then the best estimate of the market value for any home would simply be the sample mean, which is $92,069. Thus, no matter if the house has 1,500 square feet or 2,200 square feet, the best estimate of market value would still be $92,069. Because the market values vary from about $75,000 to more than $120,000, there is quite a bit of uncertainty in using the
mean as the estimate. However, from the scatter chart, we see that larger homes tend to have higher market values. Therefore, if we know that a home has 2,200 square feet, we would expect the market value estimate to be higher than that for one that has only 1,500 square feet. For example, the estimated market value of a home with 2,200 square feet would be Market Value = $32,673 + $35.036 * 2,200 = $109,752 whereas the estimated value for a home with 1,500 square feet would be Market Value = $32,673 + $35.036 * 1,500 = $85,227 The regression model explains the differences in market value as a function of the house size and provides better estimates than simply using the average of the sample data.
Using Regression Models for Prediction As we saw in Example 8.4, once we have determined the slope and intercept of the bestfitting line, we can simply substitute a value for the independent variable X in order to predict the dependent variable Y. However, there is one important caution: It is dangerous to extrapolate a regression model outside the ranges covered by the observations. For instance, if you want to predict the market value of a house that has 3,000 square feet, the results may or may not be accurate because the regression model estimates did not use any observations greater than 2,400 square feet. We cannot be sure that a linear extrapolation will hold and should not use the model to make such predictions.
320
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
8.8
Two Possible Regression Lines
▶▶ Figure
8.9
BestFitting Simple Linear Regression Line
LeastSquares Regression The mathematical basis for the bestfitting regression line is called leastsquares regression. In regression analysis, we assume that the values of the dependent variable, Y, in the sample data are drawn from some unknown population for each value of the independent variable, X. For example, in the Home Market Value data, the first and fourth observations come from a population of homes having 1,812 square feet; the second observation comes from a population of homes having 1,914 square feet; and so on. Because we are assuming that a linear relationship exists, the expected value of Y is b0 + b1X for each value of X. The coefficients b0 and b1 are population parameters that represent the intercept and slope, respectively, of the population from which a sample of
Chapter 8 Trendlines and Regression Analysis
321
observations is taken. The intercept is the mean value of Y when X = 0, and the slope is the change in the mean value of Y as X changes by one unit. Thus, for a specific value of X, we have many possible values of Y that vary around the mean. To account for this, we add an error term, e (the Greek letter epsilon), to the mean. This defines a simple linear regression model: Y = b0 + b1X + e (8.1)
However, because we don’t know the entire population, we don’t know the true values of b0 and b1. In practice, we must estimate these as best we can from the sample data. Define b0 and b1 to be estimates of b0 and b1. Thus, the estimated simple linear regression equation is Yn = b0 + b1X (8.2)
Let Xi be the value of the independent variable of the ith observation. When the value of the independent variable is Xi, then Yni = b0 + b1Xi is the estimated value of Y for Xi. One way to quantify the relationship between each point and the estimated regression equation is to measure the vertical distance between them, as illustrated in Figure 8.10. We can think of these differences, ei, as the observed errors (often called residuals) associated with estimating the value of the dependent variable using the regression line. Thus, the error associated with the ith observation is ei = Yi  Yni (8.3)
The bestfitting line should minimize some measure of these errors. Because some errors will be negative and others positive, we might take their absolute value or simply square them. Mathematically, it is easier to work with the squares of the errors. Adding the squares of the errors, we obtain the following function:
2 2 n 2 a ei = a 1Yi  Yi 2 = a 1Yi  3b0 + b1Xi 4 2 (8.4) n
n
n
i=1
i=1
i=1
If we can find the best values of the slope and intercept that minimize the sum of squares (hence the name “least squares”) of the observed errors ei, we will have found the bestfitting regression line. Note that Xi and Yi are the values of the sample data and that b0 and b1 are unknowns in equation (8.4). Using calculus, we can show that the solution that minimizes the sum of squares of the observed errors is a XiYi  nX Y n
b1 =
8.10
Measuring the Errors in a Regression Model
2 a Xi i=1
 nX
(8.5)
2
b0 = Y  b1X (8.6)
▶▶ Figure
i=1 n
Y ^ Y2 Y1 ^ Y1 Y2
2
1
X1
X2
Errors associated with individual observations
X
322
Chapter 8 Trendlines and Regression Analysis
Although the calculations for the leastsquares coefficients appear to be somewhat complicated, they can easily be performed on an Excel spreadsheet. Even better, Excel has builtin capabilities for doing this. For example, you may use the functions INTERCEPT (known_y’s, known_x’s) and SLOPE(known_y’s, known_x’s) to find the leastsquares coefficients b0 and b1. We may use the Excel function TREND(known_y’s, known_x’s, new_x’s) to estimate Y for any value of X.
EXAMPLE 8.5
Using Excel Functions to Find LeastSquares Coefficients
For the Home Market Value Excel file, the range of the dependent variable Y (market value) is C4:C45; the range of the independent variable X (square feet) is B4:B45. The Excel formula =INTERCEPT(C4:C45, B4:B45) yields b0 = 32,673 and =SLOPE(C4:C45, B4:B45) yields b1 = 35.036, as we
saw in Example 8.4. The slope tells us that for every additional square foot, the market value increases by $35.036. Using the TREND function, we may estimate the market value for a house with 1,750 square feet with the formula = TREND(C4:C45, B4:B45, 1750) = $93,987.
We could stop at this point, because we have found the bestfitting line for the observed data. However, there is a lot more to regression analysis from a statistical perspective, because we are working with sample data—and usually rather small samples—which we know have a lot of variation as compared with the full population. Therefore, it is important to understand some of the statistical properties associated with regression analysis.
Simple Linear Regression with Excel Regression analysis software tools available in Excel provide a variety of information about the statistical properties of regression analysis. The Excel Regression tool can be used for both simple and multiple linear regressions. For now, we focus on using the tool just for simple linear regression. From the Data Analysis menu in the Analysis group under the Data tab, select the Regression tool. The dialog box shown in Figure 8.11 is displayed. In the box for the Input Y Range, specify the range of the dependent variable values. In the box for the Input X Range, specify the range for the independent variable values. Check Labels if your data range contains a descriptive label (we highly recommend using this). You have the option of forcing the ▶▶ Figure
8.11
Excel Regression Tool Dialog
Chapter 8 Trendlines and Regression Analysis
323
intercept to zero by checking Constant is Zero; however, you will usually not check this box because adding an intercept term allows a better fit to the data. You also can set a Confidence Level (the default of 95% is commonly used) to provide confidence intervals for the intercept and slope parameters. In the Residuals section, you have the option of including a residuals output table by checking the boxes for Residuals, Standardized Residuals, Residual Plots, and Line Fit Plots. Residual Plots generates a chart for each independent variable versus the residual, and Line Fit Plots generates a scatter chart with the values predicted by the regression model included (however, creating a scatter chart with an added trendline is visually superior to what this tool provides). Finally, you may also choose to have Excel construct a normal probability plot for the dependent variable, which transforms the cumulative probability scale (vertical axis) so that the graph of the cumulative normal distribution is a straight line. The closer the points are to a straight line, the better the fit to a normal distribution. Figure 8.12 shows the basic regression analysis output provided by the Excel Regression tool for the Home Market Value data. The output consists of three sections: Regression Statistics (rows 3–8), ANOVA (rows 10–14), and an unlabeled section at the bottom (rows 16–18) with other statistical information. The leastsquares estimates of the slope and intercept are found in the Coefficients column in the bottom section of the output. In the Regression Statistics section, Multiple R is another name for the sample correlation coefficient, r, which was introduced in Chapter 4. Values of r range from 1 to 1, where the sign is determined by the sign of the slope of the regression line. A Multiple R value greater than 0 indicates positive correlation, that is, as the independent variable increases, the dependent variable does also; a value less than 0 indicates negative correlation—as X increases, Y decreases. A value of 0 indicates no correlation. R Square 1R2 2 is called the coefficient of determination. Earlier we noted that R2 is a measure of how well the regression line fits the data; this value is also provided by the Trendline tool. Specifically, R2 gives the proportion of variation in the dependent variable that is explained by the independent variable of the regression model. The value of R2 is between 0 and 1. A value of 1.0 indicates a perfect fit, and all data points lie on the regression line, whereas a value of 0 indicates that no relationship exists. Although we would like high values of R2, it is difficult to specify a “good” value that signifies a strong relationship because this depends on the application. For example, in scientific applications such as calibrating physical measurement equipment, R2 values close to 1 would be expected. In marketing research studies, an R2 of 0.6 or more is considered very good; however, in many social science applications, values in the neighborhood of 0.3 might be considered acceptable. ▶▶ Figure
8.12
Basic Regression Analysis Output for Home Market Value Example
324
Chapter 8 Trendlines and Regression Analysis
Adjusted R Square is a statistic that modifies the value of R2 by incorporating the sample size and the number of explanatory variables in the model. Although it does not give the actual percent of variation explained by the model as R2 does, it is useful when comparing this model with other models that include additional explanatory variables. We discuss it more fully in the context of multiple linear regression later in this chapter. Standard Error in the Excel output is the variability of the observed Yvalues from the predicted values 1Yn 2. This is formally called the standard error of the estimate, SYX. If the data are clustered close to the regression line, then the standard error will be small; the more scattered the data, the larger the standard error. EXAMPLE 8.6
Interpreting Regression Statistics for Simple Linear Regression
After running the Excel Regression tool, the first things to look for are the values of the slope and intercept, namely, the estimates b1 and b0 in the regression model. In the Home Market Value example, we see that the intercept is 32,673, and the slope (coefficient of the independent variable, square feet) is 35.036, just as we had computed earlier. In the Regression Statistics section, R 2 = 0.5347. This means that approximately 53% of the variation in market
value is explained by square feet. The remaining variation is due to other factors that were not included in the model. The standard error of the estimate is $7,287.72. If we compare this to the standard deviation of the market value, which is $10,553, we see that the variation around the regression line ($7,287.72) is less than the variation around the sample mean ($10,553). This is because the independent variable in the regression model explains some of the variation.
Regression as Analysis of Variance In Chapter 7, we introduced analysis of variance (ANOVA), which conducts an Ftest to determine whether variation due to a particular factor, such as the differences in sample means, is significantly greater than that due to error. ANOVA is commonly applied to regression to test for significance of regression. For a simple linear regression model, significance of regression is simply a hypothesis test of whether the regression coefficient b1 (slope of the independent variable) is zero:
H0: b1 = 0 H1: b1 ∙ 0 (8.7)
If we reject the null hypothesis, then we may conclude that the slope of the independent variable is not zero and, therefore, is statistically significant in the sense that it explains some of the variation of the dependent variable around the mean. Similar to our discussion in Chapter 7, you needn’t worry about the mathematical details of how F is computed, or even its value, especially since the tool does not provide the critical value for the test. What is important is the value of Significance F, which is the pvalue for the Ftest. If Significance F is less than the level of significance (typically 0.05), we would reject the null hypothesis. EXAMPLE 8.7 Interpreting Significance of Regression For the Home Market Value example, the ANOVA test is shown in rows 10–14 in Figure 8.12. Significance F, that is, the pvalue associated with the hypothesis test H0: b1 = 0 H1: b1 ∙ 0
is essentially zero (3.798 * 10 8). Therefore, assuming a level of significance of 0.05, we must reject the null hypothesis and conclude that the slope—the coefficient for square feet—is not zero. This means that home size is a statistically significant variable in explaining the variation in market value.
Chapter 8 Trendlines and Regression Analysis
325
Testing Hypotheses for Regression Coefficients Rows 17–18 of the Excel output, in addition to specifying the leastsquares coefficients, provide additional information for testing hypotheses associated with the intercept and slope. Specifically, we may test the null hypothesis that b0 or b1 equals zero. Usually, it makes little sense to test or interpret the hypothesis that b0 = 0 unless the intercept has a significant physical meaning in the context of the application. For simple linear regression, testing the null hypothesis H0: b1 = 0 is the same as the significance of regression test that we described earlier. The ttest for the slope is similar to the onesample test for the mean that we described in Chapter 7. The test statistic is
t =
b1  0 (8.8) Standard Error
and is given in the column labeled t Stat in the Excel output. Although the critical value of the tdistribution is not provided, the output does provide the pvalue for the test.
EXAMPL E 8.8
Interpreting Hypothesis Tests for Regression Coefficients
For the Home Market Value example, note that the value of t Stat is computed by dividing the coefficient by the standard error using formula (8.8). For instance, t Stat for the slope is 35.03637258>5.16738385 = 6.780292232. Because Excel does not provide the critical value with which to compare the t Stat value, we may use the pvalue to draw a conclusion. Because the pvalues for both coefficients are
essentially zero, we would conclude that neither coefficient is statistically equal to zero. Note that the pvalue associated with the test for the slope coefficient, square feet, is equal to the Significance F value. This will always be true for a regression model with one independent variable because it is the only explanatory variable. However, as we shall see, this will not be the case for multiple regression models.
Confidence Intervals for Regression Coefficients Confidence intervals (Lower 95% and Upper 95% values in the output) provide information about the unknown values of the true regression coefficients, accounting for sampling error. They tell us what we can reasonably expect to be the ranges for the population intercept and slope at a 95% confidence level. By using the confidence interval limits for the model parameters, we can determine how estimates using the model might vary. We may also use confidence intervals to test hypotheses about the regression coefficients. For example, in Figure 8.12, we see that neither confidence interval includes zero; therefore, we can conclude that b0 and b1 are statistically different from zero. Similarly, we can use them to test the hypotheses that the regression coefficients equal some value other than zero. For example, to test the hypotheses H0: b1 = B1 H1: b1 ∙ B1 we need only check whether B1 falls within the confidence interval for the slope. If it does not, then we reject the null hypothesis; otherwise, we fail to reject it.
326
Chapter 8 Trendlines and Regression Analysis
EXAMPLE 8.9
Interpreting Confidence Intervals for Regression Coefficients
For the Home Market Value data, a 95% confidence interval for the intercept is [14,823, 50,523]. Similarly, a 95% confidence interval for the slope is [24.59, 45.48]. Although the regression model is Yn = 32,673 + 35.036X, the confidence intervals suggest a bit of uncertainty about predictions using the model. Thus, although we estimated that a house with 1,750 square feet has a
market value of 32,673 + 35.036(1,750) = $93,986, if the true population parameters are at the extremes of the confidence intervals, the estimate might be as low as 14,823 + 24.59(1,750) = $57,856 or as high as 50,523 + 45.48(1,750) = $130,113. Narrower confidence intervals provide more accuracy in our predictions.
CHECK YOUR UNDERSTANDING 1. What is regression analysis? What is the difference between simple and multiple regression? 2. How can you determine whether simple linear regression would be an appropriate technique to use? 3. Why does simple linear regression provide better predictions than simply using the sample mean of the dependent variable? 4. Explain the concepts underlying leastsquares regression. 5. Explain how to interpret the results from the Excel Regression tool.
Residual Analysis and Regression Assumptions Recall that residuals are the observed errors, which are the differences between the actual values and the estimated values of the dependent variable using the regression equation. Figure 8.13 shows a portion of the residual table generated by the Excel Regression tool. The residual output includes, for each observation, the predicted value using the estimated regression equation, the residual, and the standard residual. The residual is simply the difference between the actual value of the dependent variable and the predicted value, or Yi  Yni. Figure 8.14 shows the residual plot generated by the Excel tool. This chart is actually a scatter chart of the residuals with the values of the independent variable on the xaxis. Standard residuals are residuals divided by their standard deviation. Standard residuals describe how far each residual is from its mean in units of standard deviations (similar to a zvalue for a standard normal distribution). Standard residuals are useful in checking assumptions underlying regression analysis, which we will address shortly, and to detect outliers that may bias the results. Recall that an outlier is an extreme value that is different from the rest of the data. A single outlier can make a significant difference in the regression
▶▶ Figure
8.13
Portion of Residual Output
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
327
8.14
Residual Plot for Square Feet
equation, changing the slope and intercept and, hence, how they would be interpreted and used in practice. Some consider a standardized residual outside of {2 standard deviations as an outlier. A more conservative rule of thumb would be to consider outliers outside of a {3 standard deviation range. (Commercial software packages have more sophisticated techniques for identifying outliers.)
EXAMPL E 8.10 Interpreting Residual Output For the Home Market Value data, the first observation has a market value of $90,000 and the regression model predicts $96,159.13. Thus, the residual is 90,000  96,159.13 =  $6,159.13. The standard deviation of the residuals can be computed as 7,198.299. By dividing the residual by this value, we have the standardized residual for the first observation. The value of 0.8556 tells us that the first observation is about 0.85 standard deviation below the regression line. If we check the values of all the standardized residuals, you will find that the value of the last data point is 4.53, meaning that the market value of this home, having only 1,581 square
feet, is more than 4 standard deviations above the predicted value and would clearly be identified as an outlier. (If you look back at Figure 8.7, you may have noticed that this point appears to be quite different from the rest of the data.) You might question whether this observation belongs in the data, because the house has a large value despite a relatively small size. The explanation might be an outdoor pool or an unusually large plot of land. Because this value will influence the regression results and may not be representative of the other homes in the neighborhood, you might consider dropping this observation and recomputing the regression model.
Checking Assumptions The statistical hypothesis tests associated with regression analysis are predicated on some key assumptions about the data. 1. Linearity. This is usually checked by examining a scatter diagram of the data or examining the residual plot. If the model is appropriate, then the residuals should appear to be randomly scattered about zero, with no apparent pattern. If the residuals exhibit some welldefined pattern, such as a linear trend or a parabolic shape, then there is good evidence that some other functional form might better fit the data.
328
Chapter 8 Trendlines and Regression Analysis
2. Normality of errors. Regression analysis assumes that the errors for each individual value of X are normally distributed, with a mean of zero. This can be verified either by examining a histogram of the standard residuals and inspecting for a bellshaped distribution or by using more formal goodness of fit tests. It is usually difficult to evaluate normality with small sample sizes. However, regression analysis is fairly robust against departures from normality, so in most cases, this is not a serious issue. 3. Homoscedasticity. The third assumption is homoscedasticity, which means that the variation about the regression line is constant for all values of the independent variable. This can also be evaluated by examining the residual plot and looking for large differences in the variances at different values of the independent variable. Caution should be exercised when looking at residual plots. In many applications, the model is derived from limited data, and multiple observations for different values of X are not available, making it difficult to draw definitive conclusions about homoscedasticity. If this assumption is seriously violated, then techniques other than least squares should be used for estimating the regression model. 4. Independence of errors. Finally, residuals should be independent for each value of the independent variable. For crosssectional data, this assumption is usually not a problem. However, when time is the independent variable, this is an important assumption. If successive observations appear to be correlated—for example, by becoming larger over time or exhibiting a cyclical type of pattern—then this assumption is violated. Correlation among successive observations over time is called autocorrelation and can be identified by residual plots having clusters of residuals with the same sign. Autocorrelation can be evaluated more formally using a statistical test based on a measure called the Durbin–Watson statistic. The Durbin–Watson statistic is 2 a 1ei  ei  1 2 n
D =
i=2
2 a ei i=1 n
(8.9)
This is a ratio of the squared differences in successive residuals to the sum of the squares of all residuals. D will range from 0 to 4. When successive residuals are positively autocorrelated, D will approach 0. Critical values of the statistic have been tabulated based on the sample size and number of independent variables that allow you to conclude that there is either evidence of autocorrelation or no evidence of autocorrelation or the test is inconclusive. For most practical purposes, values below 1 suggest autocorrelation; values above 1.5 and below 2.5 suggest no autocorrelation; and values above 2.5 suggest negative autocorrelation. This can become an issue when using regression in forecasting, which we discuss in the next chapter. Some software packages compute this statistic; however, Excel does not. When assumptions of regression are violated, statistical inferences drawn from the hypothesis tests may not be valid. Thus, before drawing inferences about regression models and performing hypothesis tests, these assumptions should be checked. However, other than linearity, these assumptions are not needed solely for model fitting and estimation purposes.
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
329
8.15
Histogram of Standard Residuals
EXAMPLE 8.11 Checking Regression Assumptions for the Home Market Value Data Linearity: The scatter diagram of the market value data appears to be linear; looking at the residual plot in Figure 8.14 also confirms no pattern in the residuals. Normality of errors: Figure 8.15 shows a histogram of the standard residuals for the market value data. The distribution appears to be somewhat positively skewed (particularly with the outlier) but does not appear to be a serious
departure from normality, particularly as the sample size is small. Homoscedasticity: In the residual plot in Figure 8.14, we see no serious differences in the spread of the data for different values of X, particularly if the outlier is eliminated. Independence of errors: Because the data are crosssectional, we can assume that this assumption holds.
CHECK YOUR UNDERSTANDING 1. What are standard residuals, and how can they be used? 2. Explain the assumptions behind regression analysis. 3. How can the assumptions of regression analysis be checked?
Multiple Linear Regression Many colleges try to predict student performance as a function of several characteristics. In the Excel file Colleges and Universities (see Figure 8.16), suppose that we wish to predict the graduation rate as a function of the other variables—median SAT, acceptance rate, expenditures/student, and percent in the top 10% of their high school class. It is logical to propose that schools with students who have higher SAT scores, a lower acceptance rate, a larger budget, and a higher percentage of students in the top 10% of their high school classes will tend to retain and graduate more students.
330
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
8.16
Portion of Excel File Colleges and Universities
A linear regression model with more than one independent variable is called a multiple linear regression model. Simple linear regression is just a special case of multiple linear regression. A multiple linear regression model has the form
Y = b0 + b1X1 + b2X2 + g + bkXk + e (8.10)
where Y is the dependent variable, X1,c, Xk are the independent (explanatory) variables, b0 is the intercept term, b1,c, bk are the regression coefficients for the independent variables, and e is the error term. Similar to simple linear regression, we estimate the regression coefficients—called partial regression coefficients—b0, b1, b2,c,bk, then use the model
Yn = b0 + b1X1 + b2X2 + g + bkXk (8.11)
to predict the value of the dependent variable. The partial regression coefficients represent the expected change in the dependent variable when the associated independent variable is increased by one unit while the values of all other independent variables are held constant. For the college and university data, the proposed model would be
Graduation % = b0 + b1 SAT + b2 ACCEPTANCE + b3 EXPENDITURES + b4 TOP10%HS
Thus, b2 would represent an estimate of the change in the graduation rate for a unit increase in the acceptance rate while holding all other variables constant. As with simple linear regression, multiple linear regression uses least squares to estimate the intercept and slope coefficients that minimize the sum of squared error terms over all observations. The principal assumptions discussed for simple linear regression also hold here. The Excel Regression tool can easily perform multiple linear regression; you need to specify only the full range for the independent variable data in the dialog. One caution when using the tool: The independent variables in the spreadsheet must be in contiguous columns. So you may have to manually move the columns of data around before applying the tool. The results from the Regression tool are in the same format as we saw for simple linear regression. However, some key differences exist. Multiple R and R Square (or R2) are called the multiple correlation coefficient and the coefficient of multiple determination, respectively, in the context of multiple regression. They indicate the strength of association between the dependent and independent variables. Similar to simple linear regression, R2 explains the percentage of variation in the dependent variable that is explained by the set of independent variables in the model.
Chapter 8 Trendlines and Regression Analysis
331
The interpretation of the ANOVA section is quite different from that in simple linear regression. For multiple linear regression, ANOVA tests for significance of the entire model. That is, it computes an Fstatistic for testing the hypotheses H0: b1 = b2 = g = bk = 0
H1: at least one bj is not 0
The null hypothesis states that no linear relationship exists between the dependent and any of the independent variables, whereas the alternative hypothesis states that the dependent variable has a linear relationship with at least one independent variable. If the null hypothesis is rejected, we cannot conclude that a relationship exists with every independent variable individually. The multiple linear regression output also provides information to test hypotheses about each of the individual regression coefficients. Specifically, we may test the null hypothesis that b0 (the intercept) or any bi equals zero. If we reject the null hypothesis that the slope associated with independent variable i is zero, H0: bi = 0, then we may state that independent variable i is significant in the regression model; that is, it contributes to reducing the variation in the dependent variable and improves the ability of the model to better predict the dependent variable. However, if we cannot reject H0, then that independent variable is not significant and probably should not be included in the model. We see how to use this information to identify the best model in the next section. Finally, for multiple regression models, a residual plot is generated for each independent variable. This allows you to assess the linearity and homoscedasticity assumptions of regression. EXAMPLE 8.12 Interpreting Regression Results for the Colleges and Universities Data The multiple regression results for the college and university data are shown in Figure 8.17. From the Coefficients section, we see that the model is Graduation % = 17.92 + 0.072 SAT  24.859 ACCEPTANCE  0.000136 EXPENDITURES  0.163 TOP10% HS The signs of some coefficients make sense; higher SAT scores and lower acceptance rates suggest higher graduation rates. However, we might expect that larger student expenditures and a higher percentage of top high school students would also positively influence the graduation rate. Perhaps the problem occurred because some of the best students are more demanding and change schools if their needs are not being met, some entrepreneurial students might pursue other interests before graduation, or there is sampling error. As with simple linear regression, the model should be used only for values of the independent variables within the range of the data. The value of R 2 (0.53) indicates that 53% of the variation in the dependent variable is explained by these
independent variables. This suggests that other factors not included in the model, perhaps campus living conditions, social opportunities, and so on, might also influence the graduation rate. From the ANOVA section, we may test for significance of regression. At a 5% significance level, we reject the null hypothesis because Significance F is essentially zero. Therefore, we may conclude that at least one slope is statistically different from zero. Looking at the pvalues for the independent variables in the last section, we see that all are less than 0.05; therefore, we reject the null hypothesis that each partial regression coefficient is zero and conclude that each of them is statistically significant. Figure 8.18 shows one of the residual plots from the Excel output. The assumptions appear to be met, and the other residual plots (not shown) also validate these assumptions. The normal probability plot (also not shown) does not suggest any serious departures from normality.
332
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
8.17
Multiple Regression Results for Colleges and Universities Data
▶▶ Figure
8.18
Residual Plot for Top 10% HS Variable
ANALYTICS IN PRACTICE: Using Linear Regression and Interactive Risk
Simulators to Predict Performance at Aramark4
Aramark is a leader in professional services, providing awardwinning food services, facilities management, and uniform and career apparel to health care institutions, universities and school districts, stadiums and arenas, and businesses around the world. Headquartered in Philadelphia, Aramark has approximately 255,000 employees serving clients in 22 countries.
4
Aramark’s Global Risk Management Department (GRM) needed a way to determine the statistical relationships between key business metrics (for example, employee tenure, employee engagement, a trained workforce, account tenure, service offerings) and risk metrics (for example, OSHA rate, workers’ compensation rate, customer injuries) to understand the impact of these risks on the business.
The author expresses his appreciation to John Toczek, Manager of Decision Support and Analytics at Aramark Corporation.
Chapter 8 Trendlines and Regression Analysis
Gunnar Pippel/Shutterstock
Figure 8.19 illustrates the structure of the simulators. The gauges are updated instantly as the user adjusts the sliders, showing how changes in the business environment affect the risk metrics. This visual representation made the models easy to use and understand, particularly for nontechnical employees.
Nataliia Natykach/Shutterstock
GRM sent out more than 200 surveys to multiple levels of the organization to assess the usefulness of the interactive risk simulators. One hundred percent of respondents answered “Yes” to “Were the simulators easy to use?” and 78% of respondents answered “Yes” to “Would these simulators be useful in running your business and helping you make decisions?” The deployment of interactive risk simulators to the field has been met with overwhelming positive response and recognition from leadership within all lines of business, including frontline managers, foodservice directors, district managers, and general managers.
Claudio Divizia/Shutterstock
vectorillustration/Shutterstock
GRM also needed a simple tool that field operators and the risk management team could use to predict the impact of business decisions on risk metrics before those decisions were implemented. Typical questions they would want to ask were, What would happen to our OSHA rate if we increased the percentage of part time labor? and How could we impact turnover if operations improved safety performance? Aramark maintains extensive historical data. For example, the GRM group keeps track of data such as OSHA rates, slip/trip/fall rates, injury costs, and level of compliance with safety standards; the Human Resources department monitors turnover and percentage of parttime labor; the Payroll Department keeps data on average wages; and the Training and Organizational Development Department collects data on employee engagement. Excelbased linear regression was used to determine the relationships between the dependent variables (such as OSHA rate, slip/trip/fall rate, claim cost, and turnover) and the independent variables (such as the percentage of parttime labor, average wage, employee engagement, and safety compliance). Although the regression models provided the basic analytical support that Aramark needed, the GRM team used a novel approach to implement the models for use by their clients. They developed “interactive risk simulators,” which are simple online tools that allowed users to manipulate the values of the independent variables in the regression models using interactive sliders that correspond to the business metrics and instantaneously view the values of the dependent variables (the risk metrics) on gauges similar to those found on the dashboard of a car.
Inputs: Independent Variables ▲▲ Figure
8.19
Structure of an Interactive Risk Simulator
333
Regression Models
Outputs: Dependent Variables
334
Chapter 8 Trendlines and Regression Analysis
CHECK YOUR UNDERSTANDING 1. 2. 3. 4.
What is a multiple linear regression model? How do you interpret a partial regression coefficient? What hypothesis is tested with ANOVA for a multiple linear regression model? How are ANOVA results in multiple linear regression used to test hypotheses for individual regression coefficients?
Building Good Regression Models In the colleges and universities regression example, all the independent variables were found to be significant by evaluating the pvalues of the regression analysis. This will not always be the case and leads to the question of how to build good regression models that include the “best” set of variables. Figure 8.20 shows a portion of the Excel file Banking Data, which provides data acquired from banking and census records for different ZIP codes in the bank’s current market. Such information can be useful in targeting advertising for new customers or for choosing locations for branch offices. The data show the median age of the population, median years of education, median income, median home value, median household wealth, and average bank balance. Figure 8.21 shows the results of regression analysis used to predict the average bank balance as a function of the other variables. Although the independent variables explain ▶▶ Figure
8.20
Portion of Banking Data
▶▶ Figure
8.21
Regression Analysis Results for Banking Data
Chapter 8 Trendlines and Regression Analysis
335
more than 94% of the variation in the average bank balance, you can see that at a 0.05 significance level, the pvalues indicate that both education and home value do not appear to be significant. A good regression model should include only significant independent variables. However, it is not always clear exactly what will happen when we add or remove variables from a model; variables that are (or are not) significant in one model may (or may not) be significant in another. Therefore, you should not consider dropping all insignificant variables at one time, but rather take a more structured approach. This suggests a systematic approach to building good regression models: 1. Construct a model with all available independent variables. Check for significance of the independent variables by examining the pvalues. 2. Identify the independent variable having the largest pvalue that exceeds the chosen level of significance. 3. Remove the variable identified in step 2 from the model. (Don’t remove all variables with pvalues that exceed a at the same time, but remove only one at a time.) 4. Continue until all variables are significant. In essence, this approach seeks to find a model that has only significant independent variables.
EXAMPLE 8.13 Identifying the Best Regression Model We will apply the preceding approach to the Banking Data example. The first step is to identify the variable with the largest pvalue exceeding 0.05; in this case, it is home value, and we remove it from the model and rerun the Regression tool. Figure 8.22 shows the results after removing home value. All the pvalues are now less than 0.05, so this now appears to be the best model. Notice that the
▶▶ Figure
8.22
Regression Results Without Home Value
pvalue for education, which was larger than 0.05 in the first regression analysis, dropped below 0.05 after home value was removed. This phenomenon often occurs when multicollinearity (discussed in the next section) is present and emphasizes the importance of not removing all variables with large pvalues from the original model at the same time.
336
Chapter 8 Trendlines and Regression Analysis
Adding an independent variable to a regression model will always result in R2 equal to or greater than the R2 of the original model. This is true even when the new independent variable has little true relationship with the dependent variable. Thus, trying to maximize R2 is not a useful criterion. A better way of evaluating the relative fit of different models is to use adjusted R2. Adjusted R2 reflects both the number of independent variables and the sample size and may either increase or decrease when an independent variable is added or dropped, thus providing an indication of the value of adding or removing independent variables in the model. An increase in adjusted R2 indicates that the model has improved. A criterion used to determine if a variable should be removed based on the adjusted R2 is the tstatistic. If 0 t 0 6 1, then the standard error will decrease and adjusted R2 will increase if the variable is removed. If 0 t 0 7 1, then the opposite will occur. In the banking regression results, we see that the tstatistic for home value is less than 1; therefore, we expect the adjusted R2 to increase if we remove this variable. You can follow the same iterative approach outlined before, except using tvalues instead of pvalues. One might, for example, choose the smallest ∙ t ∙ at each step. This approach will improve the adjusted R2 but may result in a model with insignificant independent variables. These approaches using the pvalues or tstatistics may involve considerable experimentation to identify the best set of variables that result in the largest adjusted R2. For large numbers of independent variables, the number of potential models can be overwhelming. For example, there are 210 = 1,024 possible models that can be developed from a set of ten independent variables. This can make it difficult to effectively screen out insignificant variables. Fortunately, automated methods—stepwise regression and best subsets—exist that facilitate this process.
Correlation and Multicollinearity As we have learned previously, correlation, a numerical value between 1 and +1, measures the linear relationship between pairs of variables. The higher the absolute value of the correlation, the greater the strength of the relationship. The sign simply indicates whether variables tend to increase together (positive) or not (negative). Therefore, examining correlations between the dependent and independent variables, which can be done using the Excel Correlation tool, can be useful in selecting variables to include in a multiple regression model because a strong correlation indicates a strong linear relationship. However, strong correlations among the independent variables can be problematic. This can potentially signify a phenomenon called multicollinearity, a condition occurring when two or more independent variables in the same regression model contain high levels of the same information and, consequently, are strongly correlated with one another and can predict each other better than the dependent variable. When significant multicollinearity is present, it becomes difficult to isolate the effect of one independent variable on the dependent variable, and the signs of coefficients may be the opposite of what they should be, making it difficult to interpret regression coefficients. Also, pvalues can be inflated, resulting in the conclusion not to reject the null hypothesis for significance of regression when it should be rejected. Some experts suggest that correlations between independent variables exceeding an absolute value of 0.7 may indicate multicollinearity. However, multicollinearity is best measured using a statistic called the variance inflation factor (VIF) for each independent variable. Sophisticated software packages usually compute these; unfortunately, Excel does not.
Chapter 8 Trendlines and Regression Analysis
337
EXAMPLE 8.14 Identifying Potential Multicollinearity Figure 8.23 shows the correlation matrix for the variables in the Colleges and Universities data. You can see that SAT and acceptance rate have moderate correlations with the dependent variable, graduation %, but the correlation between expenditures/student and top 10% HS with graduation % are relatively low. The strongest correlation, however, is between two independent variables: top 10% HS and acceptance rate. However, the value of 0.6097 does not exceed the recommended threshold of 0.7, so we can likely assume that multicollinearity is not a problem here (a more advanced analysis using VIF calculations does indeed confirm that multicollinearity does not exist).
▶▶ Figure
8.23
Correlation Matrix for Colleges and Universities Data
▶▶ Figure
8.24
Correlation Matrix for Banking Data
▶▶ Figure
8.25
Regression Results for Age, Education, and Wealth as Independent Variables
In contrast, Figure 8.24 shows the correlation matrix for all the data in the banking example. Note that large correlations exist between education and home value and also between wealth and income (in fact, the variance inflation factors do indicate significant multicollinearity). If we remove wealth from the model, the adjusted R 2 drops to 0.9201, but we discover that education is no longer significant. Dropping education and leaving only age and income in the model results in an adjusted R 2 of 0.9202. However, if we remove income from the model instead of wealth, the adjusted R 2 drops to only 0.9345, and all remaining variables (age, education, and wealth) are significant (see Figure 8.25). The R 2 value for the model with these three variables is 0.9365.
338
Chapter 8 Trendlines and Regression Analysis
Practical Issues in Trendline and Regression Modeling Finding a good regression model often requires some experimentation and trial and error. From a practical perspective, the independent variables selected should make some sense in attempting to explain the dependent variable (that is, you should have some reason to believe that changes in the independent variable will cause changes in the dependent variable even though causation cannot be proven statistically). Logic should guide your model development. In many applications, behavioral, economic, or physical theory might suggest that certain variables should belong in a model. Remember that additional variables do contribute to a higher R2 and, therefore, help to explain a larger proportion of the variation. Even though a variable with a large pvalue is not statistically significant, it could simply be the result of sampling error and a modeler might wish to keep it. Good modelers also try to have as simple a model as possible—an ageold principle known as parsimony—with the fewest number of explanatory variables that will provide an adequate interpretation of the dependent variable. In the physical and management sciences, some of the most powerful theories are the simplest. Thus, a model for the banking data that includes only age, education, and wealth is simpler than one with four variables; because of the multicollinearity issue, there would be little to gain by including income in the model. Whether the model explains 93% or 94% of the variation in bank deposits would probably make little difference. Therefore, building good regression models relies as much on experience and judgment as it does on technical analysis. An issue that one often faces in using trendlines and regression is overfitting the model. It is important to realize that sample data may have unusual variability that is different from the population; if we fit a model too closely to the sample data, we risk not fitting it well to the population in which we are interested. For instance, in fitting the crude oil prices in Example 8.2, we noted that the R2 value will increase if we fit higherorder polynomial functions to the data. While this might provide a better mathematical fit to the sample data, doing so can make it difficult to explain the phenomena rationally. The same thing can happen with multiple regression. If we add too many terms to the model, then the model may not adequately predict other values from the population. Overfitting can be mitigated by using good logic, intuition, physical or behavioral theory, and parsimony, as we have discussed. CHECK YOUR UNDERSTANDING 1. How should you properly use pvalues from multiple linear regression to build a good model? 2. Explain how to use the tstatistic to build the best multiple linear regression model. 3. What is multicollinearity, and why is it important to identify in multiple linear regression models? 4. Explain the concepts of parsimony and overfitting in building regression models.
Regression with Categorical Independent Variables Some data of interest in a regression study may be ordinal or nominal. This is common when including demographic data in marketing studies, for example. Because regression analysis requires numerical data, we could include categorical variables by coding the variables. For example, if one variable represents whether an individual is a college graduate or not, we might code no as 0 and yes as 1. Such variables are often called dummy variables.
Chapter 8 Trendlines and Regression Analysis
339
EXAMPLE 8.15 A Model with Categorical Variables The Excel file Employee Salaries, shown in Figure 8.26, provides salary and age data for 35 employees, along with an indicator of whether or not the employees have an MBA (yes or no). The MBA indicator variable is categorical; thus, we code it by replacing no with 0 and yes with 1. If we are interested in predicting salary as a function of the other variables, we would propose the model Y = b0 + b1 X 1 + b2 X 2 + e where Y = salary, X1 = age, and X2 = MBA indicator (0 or 1). After coding the MBA indicator column in the data file, we begin by running a regression on the entire data set, yielding the output shown in Figure 8.27. Note that the model explains about 95% of the variation, and the pvalues of both variables are significant. The model is Salary = 893.59 + 1,044.15 * Age + 14,767.23 * MBA
▶▶ Figure
Thus, a 30yearold with an MBA would have an estimated salary of Salary = 893.59 + 1,044.15 * 30 + 14,767.23 * 1 = $ 46,985.32 This model suggests that having an MBA increases the salary of this group of employees by almost $15,000. Note that by substituting either 0 or 1 for MBA, we obtain two models: No MBA: Salary = 893.59 + 1,044.15 * Age MBA: Salary = 15,660.82 + 1,044.15 * Age The only difference between them is the intercept. The models suggest that the rate of salary increase for age is the same for both groups. Of course, this may not be true. Individuals with MBAs might earn relatively higher salaries as they get older. In other words, the slope of age may depend on the value of MBA.
8.26
Portion of Excel File Employee Salaries
▶▶ Figure
8.27
Initial Regression Model for Employee Salaries
An interaction occurs when the effect of one variable (for example, the slope) is dependent on another variable. We can test for interactions by defining a new variable as the product of the two variables, X3 = X1 * X2, and testing whether this variable is significant, leading to an alternative model.
340
Chapter 8 Trendlines and Regression Analysis
EXAMPLE 8.16 Incorporating Interaction Terms in a Regression Model For the Employee Salaries example, we define an interaction term as the product of age 1X1 2 and MBA 1X2 2 by defining X3 = X1 * X2. The new model is Y = b0 + b1 X 1 + b2 X 2 + b3 X 3 + e In the worksheet, we need to create a new column (called Interaction) by multiplying MBA by age for each observation (see Figure 8.28). The regression results are shown in Figure 8.29. From Figure 8.29, we see that the adjusted R 2 increases; however, the pvalue for the MBA indicator variable is 0.33, indicating that this variable is not significant. We would typically drop this variable and run a regression using only age and the interaction term (Figure 8.30), and obtain the model:
▶▶ Figure
8.28
Portion of Employee Salaries Modified for Interaction Term
▶▶ Figure
8.29
Regression Results with Interaction Term
Salary = 3,323.11 + 984.25 * Age + 425.58 * MBA * Age We see that salary depends not only on whether an employee holds an MBA, but also on age, and is more realistic than the original model. However, statisticians recommend that if interactions are significant, firstorder terms should be kept in the model regardless of their pvalues. Thus, using the output in Figure 8.29, we have: Salary = 3902.51 + 971.31 * Age  2971.08 * MBA + 501.85 * MBA * Age Actually, there is little difference in the predictive ability of either model.
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
341
8.30
Regression Model for Salary Data After Dropping MBA
Categorical Variables with More Than Two Levels When a categorical variable has only two levels, as in the previous example, we coded the levels as 0 and 1 and added a new variable to the model. However, when a categorical variable has k 7 2 levels, we need to add k  1 additional variables to the model. EXAMPLE 8.17 A Regression Model with Multiple Levels of Categorical Variables The Excel file Surface Finish provides measurements of the surface finish of 35 parts produced on a lathe, along with the revolutions per minute (RPM) of the spindle and one of four types of cutting tools used (see Figure 8.31). The engineer who collected the data is interested in predicting the surface finish as a function of RPM and type of tool. Intuition might suggest defining a dummy variable for each tool type; however, doing so will cause numerical instability in the data and cause the regression tool to crash. Instead, we will need k  1 = 3 dummy variables corresponding to three of the levels of the categorical variable. The level left out will correspond to a reference, or baseline, value. Therefore, because we have k = 4 levels of tool type, we will define a regression model of the form Y = b0 + b1 X 1 + b2 X 2 + b3 X 3 + b4 X 4 + e where Y = surface finish, X1 = RPM, X2 = 1 if tool type is B and 0 if not, X3 = 1 if tool type is C and 0 if not, and X4 = 1 if tool type is D and 0 if not. Note that when X2 = X3 = X4 = 0, then, by default, the tool type is A. Substituting these values for each tool type into the model, we obtain Tool type A: Y = b0 + b1X1 + e Tool type B: Y = b0 + b1X1 + b2 + e
Tool type C: Y = b0 + b1X 1 + b3 + e Tool type D: Y = b0 + b1X1 + b4 + e For a fixed value of RPM (X1), the slopes corresponding to the dummy variables represent the difference between the surface finish using that tool type and the baseline using tool type A. To incorporate these dummy variables into the regression model, we add three columns to the data, as shown in Figure 8.32. Using these data, we obtain the regression results shown in Figure 8.33. The resulting model is Surface Finish = 24.49 + 0.098 * RPM  13.31 * Type B  20.49 * Type C  26.04 * Type D Almost 99% of the variation in surface finish is explained by the model, and all variables are significant. The models for each individual tool are Tool A: Surface Finish = 24.49 + 0.098 * RPM  13.31 * 0  20.49 * 0  26.04 * 0 = 24.49 + 0.098 * RPM Tool B: Surface Finish = 24.49 + 0.098 * RPM  13.31 * 1  20.49 * 0  26.04 * 0 = 11.18 + 0.098 * RPM
(continued )
342
Chapter 8 Trendlines and Regression Analysis
Tool C: Surface Finish = 24.49 + 0.098 * RPM  13.31 * 0  20.49 * 1  26.04 * 0 = 4.00 + 0.098 * RPM Tool D: Surface Finish = 24.49 + 0.098 * RPM  13.31 * 0  20.49 * 0  26.04 * 1 = 1.55 + 0.098 * RPM
▶▶ Figure
8.31
Portion of Excel File Surface Finish
▶▶ Figure
8.32
Data Matrix for Surface Finish with Dummy Variables
Note that the only differences among these models are the intercepts; the slopes associated with RPM are the same. This suggests that we might wish to test for interactions between the type of cutting tool and RPM; we leave this to you as an exercise.
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
343
8.33
Surface Finish Regression Model Results
CHECK YOUR UNDERSTANDING 1. Why is it necessary to code categorical independent variables in regression modeling? 2. What is an interaction? How can you test for interactions with regression? 3. Explain how to build a regression model if a categorical variable has more than two levels.
Regression Models with Nonlinear Terms Linear regression models are not appropriate for every situation. A scatter chart of the data might show a nonlinear relationship, or the residuals for a linear fit might result in a nonlinear pattern. In such cases, we might propose a nonlinear model to explain the relationship. For instance, a secondorder polynomial model would be Y = b0 + b1X + b2X 2 + e Sometimes, this is called a curvilinear regression model. In this model, b1 represents the linear effect of X on Y, and b2 represents the curvilinear effect. However, although this model appears to be quite different from ordinary linear regression models, it is still linear in the parameters (the betas, which are the unknowns that we are trying to estimate). In other words, all terms are a product of a beta coefficient and some function of the data, which are simply numerical values. In such cases, we can still apply least squares to estimate the regression coefficients. Curvilinear regression models are also often used in forecasting when the independent variable is time. This and other applications of regression in forecasting are discussed in the next chapter.
344
Chapter 8 Trendlines and Regression Analysis
EXAMPLE 8.18 Modeling Beverage Sales Using Curvilinear Regression The Excel file Beverage Sales provides data on the sales of cold beverages at a small restaurant with a large outdoor patio during the summer months (see Figure 8.34). The owner has observed that sales tend to increase on hotter days. Figure 8.35 shows linear regression results for these data. The Ushape of the residual plot (a secondorder polynomial trendline was fit to the residual data) suggests that a linear relationship is not appropriate. To apply a curvilinear regression model, add a column to the data matrix by squaring
▶▶ Figure
8.34
Portion of Excel File Beverage Sales
▶▶ Figure
8.35
Linear Regression Results for Beverage Sales
the temperatures. Now, both temperature and temperature squared are the independent variables. Figure 8.36 shows the results for the curvilinear regression model. The model is Sales = 142,850  3,643.17 * Temperature + 23.3 * Temperature2 Note that the adjusted R 2 has increased significantly from the linear model and that the residual plots now show more random patterns.
Chapter 8 Trendlines and Regression Analysis
▶▶ Figure
345
8.36
Curvilinear Regression Results for Beverage Sales
CHECK YOUR UNDERSTANDING 1. What is a curvilinear regression model? How are these models often used? 2. How can you tell if a curvilinear regression model should be used?
KEY TERMS Autocorrelation Coefficient of determination 1R2 2 Coefficient of multiple determination Curvilinear regression model Dummy variables Exponential function Homoscedasticity Interaction Leastsquares regression Linear function Logarithmic function Multicollinearity Multiple correlation coefficient
Multiple linear regression Overfitting Parsimony Partial regression coefficient Polynomial function Power function R2 (Rsquared) Regression analysis Residuals Significance of regression Simple linear regression Standard error of the estimate, SYX Standard residuals
CHAPTER 8 TECHNOLOGY HELP Useful Excel Functions INTERCEPT(known_y’s, known_x’s) Finds the intercept for a simple linear regression model.
SLOPE(known_y’s, known_x’s) Finds the slope for a simple linear regression model. TREND(known_y’s, known_x’s, new_x’s) Estimates Y for any value of X.
346
Chapter 8 Trendlines and Regression Analysis
Excel Techniques Trendline tool (Examples 8.1 and 8.2): Create a line chart or scatter chart for your data. Click the chart; this will display the Chart Tools menu. Select the Chart Tools Design tab, and then click Add Chart Element from the Chart Layouts group. From the Trendline submenu, you can select one of the options or More Trendline Options . . . . Selecting More Trendline Options will display the Format Trendline pane in the worksheet. An alternative is to right click on the data series in the chart and choose Add trendline from the popup menu. Select the radio button for the trendline. Check the boxes for Display Equation on chart and Display Rsquared value on chart. Regression tool: From the Data Analysis menu in the Analysis group under the Data tab, select the Regression tool. In the dialog box for the Input Y Range, specify the range of the dependent variable values. In the box for the Input X Range, specify the range for the independent variable values. Check Labels as appropriate. Check the optional boxes to compute and display Residuals, Standardized Residuals, Residual Plots, Line Fit Plots, and Normal Probability Plots. StatCrunch StatCrunch provides a tool for simple and multiple linear regression. You can find video tutorials and stepbystep procedures with examples at https://www.statcrunch. com/5.0/example.php. We suggest that you first view the tutorials Getting started with StatCrunch and Working with StatCrunch sessions. The following tutorial is located under the Regression and Correlation group on this Web page and explains how to perform simple linear regression. Example: Simple Linear Regression 1. Choose Stat 7 Regression 7 Simple Linear. 2. Select the X variable (independent variable) and Y variable (dependent variable) for the regression.
3. Enter an optional Where statement to specify the data rows to be included. 4. Compare results across groups by selecting an optional Group by column. 5. Click Compute! to view the results. Example: Multiple Linear Regression 1. Choose Stat 7 Regression 7 Multiple Linear. 2. Select the Y variable (dependent variable) for the regression. 3. Select the X variables (independent variables) for the regression. 4. Create interaction terms by selecting two or more variables and clicking the Add 1 or Add All button. Interaction terms will then be shown in the area to the right. To delete an interaction term, select it and click the Delete button. 5. To center the variables in each interaction term, check the Center interaction terms option. 6. Enter an optional Where statement to specify the data rows to be included. 7. Select an optional Group by column to group results. A separate regression analysis will be constructed for each distinct value of the Group by column. 8. Click Compute! to view the regression results. Analytic Solver Analytic Solver provides a set of advanced techniques for regression modeling. See the online supplement Using Best Subsets for Regression. We suggest that you first read the online supplement Getting Started with Analytic Solver Basic. This provides information for both instructors and students on how to register for and access Analytic Solver.
PROBLEMS AND EXERCISES Modeling Relationships and Trends in Data 1. Using the data in the Excel file Student Grades,
construct a scatter chart for midterm versus final exam grades and add a linear trendline. What is the model? If a student scores 75 on the midterm, what
would you predict her grade on the final exam to be? 2. A consumer products company has collected some
data relating to the advertising costs and sales of one of its products:
Chapter 8 Trendlines and Regression Analysis
9. The managing director of a consulting group has the
Advertising Cost
Sales
€300
€7,000
€350
€9,000
€400
€10,000
Overhead Costs
Billable Hours
€450
€10,600
$355,000
3,000
$400,000
4,000
$425,000
5,000
$477,000
6,000
$560,000
7,000
$580,000
8,000
following monthly data on total overhead costs and professional labor hours to bill to clients:5
What type of model would best represent the data? Use the Trendline tool to find the best among the options provided. 3. Each worksheet in the Excel file LineFit Data con
tains a set of data that describes a functional relationship between the dependent variable y and the independent variable x. Construct a line chart of each data set and use the Trendline tool to determine the bestfitting functions to model these data sets. Do not consider polynomials beyond the third degree.
a. Develop a simple linear regression model
between billable hours and overhead costs. b. Interpret the coefficients of your regression model. Specifically, what does the fixed component of the model mean to the consulting firm?
4. Using the data in the Excel file Kidney, determine
the relationship between frailty rates in people with chronic kidney disease and the age of a person by constructing a scatter chart for the data. Find the bestfitting function using the Trendline tool. Interpret your findings.
c. If a special job was available requiring 4,500
billable hours that would contribute a margin of $240,000 before overhead, would the job be attractive? 10. Using the data in the Excel file Kidney, apply the
Excel Regression tool using kidney frailty rates and age of a person as the variables. Interpret all key regression results, hypothesis tests, and confidence intervals in the output.
5. Use the data in the Excel file MBA Motivation and
Salary Expectations to determine how well a linear function models the relationship between pre and postMBA salaries. Compare it to a thirdorder polynomial. Would the polynomial model have a logical explanation?
11. The Excel file Horsepower versus Gasoline provides
data on the amount of gasoline used per 100 miles by different brands of cars along with their horsepower. a. Construct a scatter diagram for gasoline used per 100 miles and horsepower in the given Excel file. Determine the relationship between the variables. b. Use the Excel Regression tool to develop a model for predicting the amount of gasoline used per 100 miles. Explain the statistical significance of the model and the R2 value.
6. Using the data in the Excel file Weddings, construct
scatter charts to determine whether any linear relationship appears to exist between (1) the wedding cost (X) and attendance (Y), (2) the wedding cost (X) and the value rating (Y), and (3) the couple’s income (X) and wedding cost (Y) only for the weddings paid for by the bride and groom. Then find the bestfitting functions using the Trendline tool for each of these charts. Simple Linear Regression
347
12. Using the data in the Excel file Student Grades, apply
the Excel Regression tool using the midterm grade as the independent variable and the final exam grade as the dependent variable. Interpret all key regression results, hypothesis tests, and confidence intervals in the output.
7. Using the results of fitting the Home Market Value
regression line in Example 8.4, compute the errors associated with each observation using formula (8.3) and construct a frequency distribution and histogram. 8. Set up an Excel worksheet to apply formulas (8.5)
and (8.6) to compute the values of b0 and b1 for the data in the Excel file Home Market Value and verify that you obtain the same values as in Examples 8.4 and 8.5.
5
Modified from Charles T. Horngren, George Foster, and Srikant M. Datar, Cost Accounting: A Managerial Emphasis, 9th ed. (Englewood Cliffs, NJ: Prentice Hall, 1997): 371.
348
Chapter 8 Trendlines and Regression Analysis
13. Using the Excel file Weddings, apply the Regression
19. Use the results for Problem 13 (Weddings) to analyze
tool using the wedding cost as the dependent variable and attendance as the independent variable. a. What is the regression model? b. Interpret all key regression results, hypothesis tests, and confidence intervals in the output. c. If a couple is planning a wedding for 175 guests, how much should they budget?
the residuals to determine if the assumptions underlying the regression analysis are valid. In addition, use the standard residuals to determine if any possible outliers exist. Multiple Linear Regression 20. The Excel file World Tour Sales provides data on
Regression tool using the wedding cost as the dependent variable and the couple’s income as the independent variable, only for those weddings paid for by the bride and groom. Interpret all key regression results, hypothesis tests, and confidence intervals in the output.
sales (£) and the number of radio, TV, and newspaper ads promoting the world tour for a group of countries. Develop simple linear regression models for predicting sales as a function of the number of each type of ad. Compare these results to a multiple linear regression model using both independent variables. State each model and explain RSquare, Significance F, and pvalues.
15. A deepfoundation engineering contractor has bid
21. Using the data in the Excel file Home Market Value,
14. Using the Excel file Weddings, apply the Excel
on a foundation system for a new building housing the world headquarters for a Fortune 500 company. A part of the project consists of installing 311 auger cast piles. The contractor was given bid information for costestimating purposes, which consisted of the estimated depth of each pile; however, actual drill footage of each pile could not be determined exactly until construction was performed. The Excel file Pile Foundation contains the estimates and actual pile lengths after the project was completed. a. Construct a scatter chart and interpret it. b. Develop a linear regression model to estimate the actual pile length as a function of the estimated pile lengths. What do you conclude? Residual Analysis and Regression Assumptions 16. Use the results for Problem 10 (Demographics) to
analyze the residuals to determine if the assumptions underlying the regression analysis are valid. In addition, use the standard residuals to determine if any possible outliers exist. 17. Use the results in Problem 11 (Horsepower versus
Gasoline) to analyze the residuals to determine if the assumptions underlying the regression analysis are valid. In addition, use the standard residuals to determine if any possible outliers exist. 18. Use the results for Problem 12 (Student Grades) to
analyze the residuals to determine if the assumptions underlying the regression analysis are valid. In addition, use the standard residuals to determine if any possible outliers exist.
develop a multiple linear regression model for estimating the market value as a function of both the age and size of the house. State the model and explain R2, Significance F, and pvalues. 22. Use the data in the Excel file Canned Food to find a
multiple regression model to predict the total sales as a function of all the other given variables. State the model and explain R2, Significance F, and pvalues. 23. Use the data in the Excel file Neighborhoods to find
a multiple regression model to predict the total sales as a function of all the other given variables. State the model and explain R2, Significance F, and pvalues. Building Good Regression Models 24. Using the data in the Excel file Credit Card Spend
ing, develop a multiple linear regression model for estimating the average credit card expenditure as a function of both the income and family size. Predict the average expense of a family that has two members and an income of $188,000 per annum, and another that has three members and an income of $39,000 income per annum. 25. Use the data in the Excel file EU Cost of Living Adjust
ments to find the best multiple regression model to predict the salary as a function of the adjusted cost of living rates. What would the comparable salary be for a country in the European Union with the following adjustments: groceries: 5%; housing: 8%; utilities: −3%; transportation: 1%; and health care: 9%? 26. Use the data in the Excel file Job Satisfaction to find
the best multiple regression model to predict the
Chapter 8 Trendlines and Regression Analysis
overall satisfaction as a function of the other variables. What managerial implications does your result have? 27. The Excel file Breakfast Foods provides a variety of
nutritional information about 67 breakfastfood items in supermarkets in different parts of the world. Use regression analysis to find the best model that explains the relationship between calories and the other variables. Keep in mind the principle of parsimony. 28. The Excel file Salary Data provides information on
349
32. The Excel file Golfing Statistics provides data for a
portion of the 2010 professional season for the top 25 golfers. a. Find the best multiple regression model for predicting earnings/event as a function of the other variables. b. Find the best multiple regression model for predicting average score as a function of GIR, putts/ round, driving accuracy, and driving distance.
current salary, beginning salary, previous experience (in months) when hired, and total years of education for a sample of 100 employees in a firm. a. Find the multiple regression model for predicting current salary as a function of the other variables.
33. Use the pvalue criterion to find the best model for
b. Find the best model for predicting current salary
34. A researcher is analyzing the body fat percentage
using the tvalue criterion. 29. The Excel file Credit Approval Decisions provides
information on credit history for a sample of banking customers. Use regression analysis to identify the best model for predicting the credit score as a function of the other numerical variables, using both the pvalue and tstatistic criteria. How do the models compare? Which would you choose? 30. The Excel file Women Height Data provides infor
mation on women’s height, their mother’s height, their father’s height, amount of protein intake (per day), and the hours of exercise (per day) for a select sample of women. Use the Correlation tool to find the correlation matrix for all variables to justify the potential of multicollinearity. Then identify the best regression model for predicting the women’s height using the pvalue criteria. 31. The Excel file Major League Baseball provides data
predicting the number of points scored per game by football teams using the data in the Excel file National Football League. Does the model make logical sense? based on the measurements of triceps, thigh, and midarm of a person (in inches). The data is recorded in the Excel file Body Fat Percentage. a. Use the Correlation tool to develop the correlation matrix and make a conclusion for it. Then use the Regression tool for this set of variables to develop a multiple regression model, and identify which independent variable is not appropriate in the model. b. Remove the inappropriate independent in part a in the multiple regression model and rerun the process using the Regression tool until the best model is found. c. Use the best multiple regression model to predict the body fat percentage given the following measurements: 25.1 inches of triceps, 50.78 inches of thigh, and 33.56 inches of midarm.
for one season.
Regression with Categorical Independent Variables
a. Construct and examine the correlation matrix. Is
35. A leading car manufacturer tracks the data of its used
multicollinearity a potential problem? b. Suggest an appropriate set of independent variables that predict the number of wins by examining the correlation matrix. c. Find the best multiple regression model for predicting the number of wins having only significant independent variables. How good is your model? Does it use the same variables you thought were appropriate in part b?
cars for resale. The Excel file Car Sales contains information on the selling price of the car, fuel type (diesel or petrol), horsepower (HP), and manufacture year. a. Develop a multiple linear regression model for the selling price as a function of fuel type and HP without any interaction term. b. Determine if any interaction exists between fuel type and HP and find the best model.
350
Chapter 8 Trendlines and Regression Analysis
c. What is the predicted price for either a petrol or
Units Produced
diesel car with a horsepower of 69? 36. For the Car Sales data described in Problem 35,
develop a regression model for selling price as a function of horsepower and manufacture year, incorporating an interaction term. What would be the predicted price for a car manufactured in either 2002 or 2003 with a horsepower of 69? How do these predictions compare to the overall average price in each year? 37. For the Excel file Job Satisfaction, develop a regres
sion model for overall satisfaction as a function of years of service and department that has the largest R2. (Note that the categorical variable department has multiple levels and will require the use of multiple dummy variables, similar to Example 8.17.) Which department, if any, has the highest impact on satisfaction?
Cost
500
$12,500
1,000
$25,000
1,500
$32,500
2,000
$40,000
2,500
$45,000
3,000
$50,000
39. A product manufacturer wishes to determine the rela
tionship between the shelf space of the product and its sales. Past data indicates the following sales and shelf space in its stores: Sales
Shelf Space
$25,000
5 square feet
$15,000
3.2square feet
$28,000
5.4square feet
$30,000
6.1square feet
$17,000
4.3square feet
Regression Models with Nonlinear Terms
$16,000
3.1square feet
38. Cost functions are often nonlinear with volume
$12,000
2.6 square feet
because production facilities are often able to produce larger quantities at lower rates than smaller quantities.6 Using the following data, apply simple linear regression and examine the residual plot. What do you conclude? Construct a scatter chart and use the Trendline feature to identify the best type of curvilinear trendline that maximizes R2.
$21,000
6.4 square feet
$19,000
4.9 square feet
$27,000
5.7 square feet
Using these data points, apply simple linear regression and examine the residual plot. What do you conclude? Construct a scatter chart and use the Trendline tool to identify the best type of curvilinear trendline that maximizes R2.
CASE: PERFORMANCE LAWN EQUIPMENT In reviewing the data in the Performance Lawn Equipment Database, Elizabeth Burke noticed that defects received from suppliers have decreased (worksheet Defects After Delivery). Upon investigation, she learned that in 2014, PLE experienced some quality problems due to an increasing number of defects in materials received from suppliers. The company instituted an initiative in August 2015 to work with suppliers to reduce these defects, to more closely coordinate deliveries, and to improve materials quality through reengineering supplier production policies. Ms. Burke noted that the program appeared to reverse an 6 Horngren, 7
increasing trend in defects; she would like to predict what might have happened had the supplier initiative not been implemented and how the number of defects might further be reduced in the near future. In meeting with PLE’s human resources director, Ms. Burke also discovered a concern about the high rate of turnover in its field service staff. Senior managers have suggested that the department look closer at its recruiting policies, particularly to try to identify the characteristics of individuals that lead to greater retention. However, in a recent staff meeting, HR managers could not agree on these characteristics.
Foster, and Datar, Cost Accounting: A Managerial Emphasis, 9th ed.: 349. Horngren, Foster, and Datar, Cost Accounting: A Managerial Emphasis, 9th ed.: 349.
Some argued that years of education and grade point averages were good predictors. Others argued that hiring more mature applicants would lead to greater retention. To study these factors, the staff agreed to conduct a statistical study to determine the effect that years of education, college grade point average, and age when hired have on retention. A sample of 40 field service engineers hired ten years ago was selected to determine the influence of these variables on how long each individual stayed with the company. Data are compiled in the Employee Retention worksheet. Finally, as part of its efforts to remain competitive, PLE tries to keep up with the latest in production technology. This is especially important in the highly competitive lawn mower line, where competitors can gain a real advantage if they develop more costeffective means of production. The lawn mower division therefore spends a great deal of effort in testing new technology. When new
Chapter 8 Trendlines and Regression Analysis
351
production technology is introduced, firms often experience learning, resulting in a gradual decrease in the time required to produce successive units. Generally, the rate of improvement declines until the production time levels off. One example is the production of a new design for lawn mower engines. To determine the time required to produce these engines, PLE produced 50 units on its production line; test results are given on the worksheet Engines in the database. Because PLE is continually developing new technology, understanding the rate of learning can be useful in estimating future production costs without having to run extensive prototype trials, and Ms. Burke would like a better handle on this. Use trendlines and regression analysis to assist her in evaluating the data in these three worksheets and reaching useful conclusions. Summarize your work in a formal report with all appropriate results and analyses.
This page intentionally left blank
CHAPTER
Forecasting Techniques
9
rawpixel/123RF
LEARNING OBJECTIVES ■■
■■ ■■
■■
■■
After studying this chapter, you will be able to:
Explain how judgmental approaches are used for forecasting. List different types of statistical forecasting models. Apply moving average and exponential smoothing models to stationary time series. State three error metrics used for measuring forecast accuracy and explain the differences among them. Apply double exponential smoothing models to time series with a linear trend.
■■
■■
■■
■■
Use HoltWinters and regression models to forecast time series with seasonality. Apply HoltWinters forecasting models to time series with both trend and seasonality. Identify the appropriate choice of forecasting model based on the characteristics of a time series. Explain how regression techniques can be used to forecast with explanatory or causal variables.
354
Chapter 9 Forecasting Techniques
Managers require good forecasts of future events to make good decisions. For example, forecasts of interest rates, energy prices, and other economic indicators are needed for financial planning; sales forecasts are needed to plan production and workforce capacity; and forecasts of trends in demographics, consumer behavior, and technological innovation are needed for longterm strategic planning. The government also invests significant resources on predicting shortrun U.S. business performance using the index of leading economic indicators (LEI). This index focuses on the performance of individual businesses, which often is highly correlated with the performance of the overall economy and is used to forecast economic trends for the nation as a whole. In this chapter, we introduce some common methods and approaches to forecasting, including both qualitative and quantitative techniques. Business analysts may choose from a wide range of forecasting techniques to support decision making. Selecting the appropriate method depends on the characteristics of the forecasting problem, such as the time horizon of the variable being forecast, as well as available information on which the forecast will be based. Three major categories of forecasting approaches are qualitative and judgmental techniques, statistical timeseries models, and explanatory/causal methods. In this chapter, we introduce a variety of forecasting techniques and use basic Excel tools and linear regression to implement them on spreadsheets.
ANALYTICS IN PRACTICE: Forecasting CallCenter Demand at
L.L.Bean1
Many of you are familiar with L.L.Bean, a retailer of highquality outdoor gear. A large percentage of the company’s sales are generated through orders to its call center (the call center can account for over 70% of the total sales volume). Calls to the L.L.Bean call center are classified into two types: telemarketing (TM) and telephone inquiry (TI). TM calls are those made for placing an order, whereas TI calls involve customer inquiries, such as order status or order problems. TM calls and TI calls differ in duration and volume. The annual call volume for TM calls is much higher than that for TI calls, but the duration of a TI call is generally much longer than the duration of a TM call.
1
Accurately forecasting the demand of TM and TI calls is very important to L.L.Bean to reduce costs. Accurate forecasts allow for properly planning the number of agents to have on hand at any point in time. Too few agents result in lost sales, loss of customer loyalty, excessive queue times, and increased phone charges. Too many agents obviously result in unnecessary labor costs. L.L.Bean developed analytical forecasting models for both TM and TI calls. These models took into account historical trends, seasonal factors, and external explanatory variables such as holidays and catalog mailings. The estimated benefit from better precision from the two forecasting models was approximately $300,000 per year.
Andrews, B.H., and S. M. Cunningham, “L.L. Bean Improves CallCenter Forecasting,” Interfaces, Vol. 25, No. 6, pp.113, NovemberDecember, 1995.
Chapter 9 Forecasting Techniques
355
Qualitative and Judgmental Forecasting Qualitative and judgmental techniques rely on experience and intuition; they are necessary when historical data are not available or when the decision maker needs to forecast far into the future. For example, a forecast of when the next generation of a microprocessor will be available and what capabilities it might have will depend greatly on the opinions and expertise of individuals who understand the technology. Another use of judgmental methods is to incorporate nonquantitative information, such as the impact of government regulations or competitor behavior, in a quantitative forecast. Judgmental techniques range from such simple methods as a manager’s opinion or a groupbased jury of executive opinion to more structured approaches such as historical analogy and the Delphi method.
Historical Analogy One judgmental approach is historical analogy, in which a forecast is obtained through a comparative analysis with a previous situation. For example, if a new product is being introduced, the response of consumers to marketing campaigns to similar, previous products can be used as a basis to predict how the new marketing campaign might fare. Of course, temporal changes or other unique factors might not be fully considered in such an approach. However, a great deal of insight can often be gained through an analysis of past experiences.
EXAMPLE 9.1
Predicting the Price of Oil
In early 1998, the price of oil was about $22 a barrel. However, in mid1998, the price of a barrel of oil dropped to around $11. The reasons for this price drop included an oversupply of oil from new production in the Caspian Sea region, high production in nonOPEC regions, and lowerthannormal demand. In similar circumstances in the past, OPEC would meet and take action to raise the price of
oil. Thus, from historical analogy, we might forecast a rise in the price of oil. OPEC members did, in fact, meet in mid1998 and agreed to cut their production, but nobody believed that they would actually cooperate effectively, and the price continued to drop for a time. Subsequently, in 2000, the price of oil rose dramatically, falling again in late 2001.
Analogies often provide good forecasts, but you need to be careful to recognize new or different circumstances. Another analogy is international conflict relative to the price of oil. Should war break out, the price would be expected to rise, analogous to what it has done in the past.
The Delphi Method A popular judgmental forecasting approach, called the Delphi method, uses a panel of experts, whose identities are typically kept confidential from one another, to respond to a sequence of questionnaires. After each round of responses, individual opinions, edited to ensure anonymity, are shared, allowing each to see what the other experts think. Seeing other experts’ opinions helps to reinforce those in agreement and to influence those who did not agree to possibly consider other factors. In the next round, the experts revise their estimates, and the process is repeated, usually for no more than two or three rounds. The Delphi method promotes unbiased exchanges of ideas and discussion and usually results in
356
Chapter 9 Forecasting Techniques
some convergence of opinion. It is one of the better approaches to forecasting longrange trends and impacts.
Indicators and Indexes Indicators and indexes generally play an important role in developing judgmental forecasts. Indicators are measures that are believed to influence the behavior of a variable we wish to forecast. By monitoring changes in indicators, we expect to gain insight about the future behavior of the variable to help forecast the future.
EXAMPLE 9.2
Economic Indicators
One variable that is important to the nation’s economy is the gross domestic product (GDP), which is a measure of the value of all goods and services produced in the United States. Despite its shortcomings (for instance, unpaid work such as housekeeping and child care is not measured; production of poorquality output inflates the measure, as does work expended on corrective action), it is a practical and useful measure of economic performance. Like most time series, the GDP rises and falls in a cyclical fashion. Predicting future trends in the GDP is often done by analyzing leading indicators—series that tend to rise and fall for some
predictable length of time prior to the peaks and valleys of the GDP. One example of a leading indicator is the formation of business enterprises; if the rate of new businesses grows, we would expect the GDP to increase in the near future. Additional examples of leading indicators are the percent change in the money supply (M1) and net change in business loans. Other indicators, called lagging indicators, tend to have peaks and valleys that follow those of the GDP. Some lagging indicators are the Consumer Price Index, prime rate, business investment expenditures, and inventories on hand. The GDP can be used to predict future trends in these indicators.
Indicators are often combined quantitatively into an index, a single measure that weights multiple indicators, thus providing a measure of overall expectation. For example, financial analysts use the Dow Jones Industrial Average as an index of general stock market performance. Indexes do not provide a complete forecast but rather a better picture of direction of change and thus play an important role in judgmental forecasting.
EXAMPLE 9.3
Leading Economic Indicators
The Department of Commerce initiated an index of leading indicators to help predict future economic performance. Components of the index include the following: ■■ average weekly hours, manufacturing ■■ average weekly initial claims, unemployment
insurance ■■ new orders, consumer goods, materials ■■ vendor performance—slower deliveries ■■ new orders, nondefense capital goods ■■ building permits, private housing ■■ stock prices, 500 common stocks (Standard & Poor) ■■ money supply
■■ interest rate spread ■■ index of consumer expectations (University of Michigan)
Business Conditions Digest included more than 100 time series in seven economic areas. This publication was discontinued in March 1990, but information related to the index of leading indicators was continued in Survey of Current Business. In December 1995, the U.S. Department of Commerce sold this data source to the Conference Board, which now markets the information under the title Business Cycle Indicators; information can be obtained at its Web site (www.conferenceboard.org). The site includes excellent current information about the calculation of the index as well as its current components.
Chapter 9 Forecasting Techniques
357
CHECK YOUR UNDERSTANDING 1. Explain how historical analogy and the Delphi method are used for judgmental forecasting. 2. Define an indicator and index. How are they used in forecasting?
Statistical Forecasting Models Statistical timeseries models find greater applicability for shortrange forecasting problems. A time series is a stream of historical data, such as weekly sales. We characterize the values of a time series over T periods as At, t = 1, 2, c, T. Timeseries models assume that whatever forces have influenced sales in the recent past will continue into the near future; thus, forecasts are developed by extrapolating these data into the future. Time series generally have one or more of the following components: random behavior, trends, seasonal effects, and cyclical effects. Time series that do not have trend, seasonal, or cyclical effects but are relatively constant and exhibit only random behavior are called stationary time series. Many forecasts are based on analysis of historical timeseries data and are predicated on the assumption that the future is an extrapolation of the past. A trend is a gradual upward or downward movement of a time series over time. EXAMPLE 9.4
Identifying Trends in a Time Series
Figure 9.1 shows a chart of total energy consumption from the data in the Excel file Energy Production & Consumption. This time series shows an upward trend. However, we see that energy consumption was rising quite rapidly in a linear fashion during the 1960s, then
leveled off for a while and began increasing at a slower rate through the 1980s and 1990s. At the end of the time series, we actually see a slight downward trend. This time series, then, is composed of several different short trends.
Time series may also exhibit shortterm seasonal effects well as longerterm cyclical effects, or nonlinear trends. A seasonal effect is one that repeats at fixed intervals of time, typically a year, month, week, or day. At a neighborhood grocery store, for instance, shortterm seasonal patterns may occur over a week, with the heaviest volume of customers on weekends; seasonal patterns may also be evident during the course of a day, with higher volumes in the mornings and late afternoons. Figure 9.2 shows seasonal changes in natural gas usage for a homeowner over the course of a year (Excel file Gas & Electric). Cyclical effects describe ups and downs over a much longer time frame, such as several years. Figure 9.3 shows a chart of the data in the Excel file Federal Funds Rates. We see some evidence of longterm cycles in the time series driven by economic factors, such as periods of inflation and recession. Although visual inspection of a time series to identify trend, seasonal, or cyclical effects may work in a naïve fashion, such unscientific approaches may be a bit unsettling to a manager making important decisions. Subtle effects and interactions of seasonal and cyclical factors may not be evident from simple visual extrapolation of data. Statistical methods, which involve more formal analyses of time series, are invaluable in developing good forecasts. A variety of statistically based forecasting methods for time series
358
Chapter 9 Forecasting Techniques
▶▶ Figure
9.1
Total Energy Consumption Time Series
▶▶ Figure
9.2
Seasonal Effects in Natural Gas Usage
▶▶ Figure
9.3
Cyclical Effects in Federal Funds Rates
are commonly used. Among the most popular are moving average methods, exponential smoothing, and regression analysis. These can be implemented very easily on a spreadsheet using basic functions and Data Analysis tools available in Microsoft Excel, as well as with more powerful software such as Analytic Solver (see the online supplement described in the Technology Help section). Moving average and exponential smoothing models work best for time series that do not exhibit trends or seasonal factors. For time series that involve trends and/or seasonal factors, other techniques have been developed. These include double moving average and exponential smoothing models, seasonal additive and multiplicative models, and HoltWinters additive and multiplicative models.
Chapter 9 Forecasting Techniques
359
CHECK YOUR UNDERSTANDING 1. What is a time series? 2. What components often comprise a time series? 3. Explain the difference between seasonal and cyclical effects in a time series.
Forecasting Models for Stationary Time Series Two simple approaches that are useful over short time periods when trend, seasonal, or cyclical effects are not significant are moving average and exponential smoothing models.
Moving Average Models The simple moving average method is a smoothing method based on the idea of averaging random fluctuations in the time series to identify the underlying direction in which the time series is changing. Because the moving average method assumes that future observations will be similar to the recent past, it is most useful as a shortrange forecasting method. Although this method is very simple, it has proven to be quite useful in stable environments, such as inventory management, in which it is necessary to develop forecasts for a large number of items. Specifically, the simple moving average forecast for the next period is computed as the average of the most recent k observations. Specifically, let At represent the observed value in period t, and Ft the forecast for period t. Then the forecast for the period t + 1 is computed as
Ft + 1 =
At + At1 + c + Atk + 1 (9.1) k
The value of k is somewhat arbitrary, although its choice affects the accuracy of the forecast. The larger the value of k, the more the current forecast is dependent on older data, and the forecast will not react as quickly to fluctuations in the time series. The smaller the value of k, the quicker the forecast responds to changes in the time series. Also, when k is larger, extreme values have less effect on the forecasts. (In the next section, we discuss how to select k by examining errors associated with different values.)
EXAMPLE 9.5
Moving Average Forecasting
The Excel file Tablet Computer Sales contains data for the number of units sold for the past 17 weeks. Figure 9.4 shows a chart of these data. The time series appears to be relatively stable, without trend, seasonal, or cyclical effects; thus, a moving average model would be appropriate. The observed values for weeks 15, 16, and 17 are A15 = 82, A16 = 71, and A17 = 50. Setting k = 3, and using equation (9.1), the threeperiod moving average forecast for week 18 is F18 =
(A17 + A16 + A15) 82 + 71 + 50 = = 67.67 3 3
Moving average forecasts can be generated easily on a spreadsheet. Figure 9.5 shows the computations for a threeperiod moving average forecast of tablet computer sales. You should verify these by hand to confirm your understanding. Figure 9.6 shows a chart that contrasts the data with the forecasted values. Although the forecast tracks the actual data fairly well, notice that the forecast overestimates the actual value when the trend is down, while it underestimates the actual value when the trend is up. This is because it uses past data, and thus lags the changes in the data.
360
Chapter 9 Forecasting Techniques
▶▶ Figure
9.4
Chart of Weekly Tablet Computer Sales
▶▶ Figure
9.5
Excel Implementation of Moving Average Forecast
▶▶ Figure
9.6
Chart of Units Sold and Moving Average Forecast
Chapter 9 Forecasting Techniques
361
Moving average forecasts can also be obtained from Excel’s Data Analysis options. EXAMPLE 9.6
Using Excel’s Moving Average Tool
For the Tablet Computer Sales Excel file, select Data Analysis and then Moving Average from the Analysis group. Excel displays the dialog box shown in Figure 9.7. You need to enter the Input Range of the data, the Interval (the value of k), and the first cell of the Output Range. To align the actual data with the forecasted values in the worksheet, select the first cell of the Output Range to be one row below the first row of the time series. (See Figure 9.8. The first value of the time series starts in cell B4, so the output range is chosen to start in cell C5.). You may also obtain a chart of the data and the moving averages, as well as a column of standard errors, by checking
the appropriate boxes. However, we do not recommend using the chart or error options because the forecasts generated by this tool are not properly aligned with the data (the forecast value aligned with a particular data point represents the forecast for the next month) and, thus, can be misleading. Rather, we recommend that you generate your own chart, as we did in Figure 9.6. Figure 9.8 shows the results produced by the Moving Average tool (with some customization of the formatting). Note that the forecast for week 18 is aligned with the actual value for week 17 on the chart. Compare this to Figure 9.6 and you can see the difference.
Error Metrics and Forecast Accuracy The quality of a forecast depends on how accurate it is in predicting future values of a time series. In the simple moving average model, different values for k will produce different forecasts. How do we know which is the best value for k? The error, or residual, in a forecast is the difference between the forecast and the actual value of the time series (once it is known). In Figure 9.6, the forecast error is simply the vertical distance between the forecast and the data for the same time period. To analyze the effectiveness of different forecasting models, we can define error metrics, which compare quantitatively the forecast with the actual observations. Three metrics that are commonly used are the mean absolute deviation, mean square error, and mean absolute percentage error. The mean absolute deviation (MAD) is the absolute difference between the actual value and the forecast, averaged over a range of forecasted values: a ∙ At  Ft ∙ n
MAD =
t=1
n
(9.2)
where At is the actual value of the time series at time t, Ft is the forecast value for time t, and n is the number of forecast values (not the number of data points since we do not have ▶▶ Figure
9.7
Excel Moving Average Tool Dialog
362
Chapter 9 Forecasting Techniques
▶▶ Figure
9.8
Results of Excel Moving Average Tool (Note misalignment of forecasts with actual sales in the chart.)
a forecast value associated with the first k data points). MAD provides a robust measure of error and is less affected by extreme observations. Mean square error (MSE) is probably the most commonly used error metric. It penalizes larger errors because squaring larger numbers has a greater impact than squaring smaller numbers. The formula for MSE is 2 a 1At  Ft 2 n
t=1
MSE =
n
(9.3)
Again, n represents the number of forecast values used in computing the average. Sometimes the square root of MSE, called the root mean square error (RMSE), is used: 2 a 1At  Ft 2 n
RMSE =
H
t=1
n
(9.4)
Note that unlike MSE, RMSE is expressed in the same units as the data (similar to the difference between a standard deviation and a variance), allowing for more practical comparisons. A fourth commonly used metric is mean absolute percentage error (MAPE). MAPE is the average of absolute errors divided by actual observation values. a n
MAPE =
t=1
2
At  Ft 2 At * 100 (9.5) n
The values of MAD and MSE depend on the measurement scale of the timeseries data. For example, forecasting profit in the range of millions of dollars would result in very large MAD and MSE values, even for very accurate forecasting models. On the other hand, market share is measured in proportions; therefore, even bad forecasting models will have small values of MAD and MSE. Thus, these measures have no meaning except in comparison with other models used to forecast the same data. Generally, MAD is less affected by
Chapter 9 Forecasting Techniques
363
extreme observations and is preferable to MSE if such extreme observations are considered rare events with no special meaning. MAPE is different in that the measurement scale is eliminated by dividing the absolute error by the timeseries data value. This allows a better relative comparison. Although these comments provide some guidelines, there is no universal agreement on which measure is best. EXAMPLE 9.7
Using Error Metrics to Compare Moving Average Forecasts
The metrics we have described can be used to compare different moving average forecasts for the Tablet Computer Sales data. A spreadsheet that shows the forecasts as well as the calculations of the error metrics for two, three, and fourperiod moving average models is given in Figure 9.9. The error is the difference between the actual value of the units sold and the forecast. To compute MAD, we first compute the absolute values of the errors and
then average them. For MSE, we compute the squared errors and then find the average (we may easily compute RMSE if desired). For MAPE, we find the absolute values of the errors divided by the actual observation multiplied by 100 and then average them. The results suggest that a twoperiod moving average model provides the best forecast among these alternatives because the error metrics are all smaller than for the other models.
Exponential Smoothing Models A versatile, yet highly effective, approach for shortrange forecasting is simple exponential smoothing. The basic simple exponential smoothing model is
Ft + 1 = 11  a2Ft + aAt
= Ft + a1At  Ft 2 (9.6)
where Ft + 1 is the forecast for time period t + 1, Ft is the forecast for period t, At is the observed value in period t, and a is a constant between 0 and 1 called the smoothing constant. To begin, set F1 and F2 equal to the actual observation in period 1, A1. Using the two forms of the forecast equation just given, we can interpret the simple exponential smoothing model in two ways. In the first model, the forecast for the next period, Ft + 1, is a weighted average of the forecast made for period t, Ft, and the actual
▲▲ Figure
9.9
Error Metrics Alternative Moving Average Forecasts
364
Chapter 9 Forecasting Techniques
observation in period t, At. The second form of the model, obtained by simply rearranging terms, states that the forecast for the next period, Ft + 1, equals the forecast for the last period, Ft, plus a fraction a of the forecast error made in period t, At  Ft. Thus, to make a forecast once we have selected the smoothing constant, we need to know only the previous forecast and the actual value. By repeated substitution for Ft in the equation, it is easy to demonstrate that Ft + 1 is a decreasingly weighted average of all past timeseries data. Thus, the forecast actually reflects all the data, provided that a is strictly between 0 and 1. EXAMPLE 9.8
Using Exponential Smoothing to Forecast Tablet Computer Sales
For the Tablet Computer Sales data, the forecast for week 2 is 88, the actual observation for week 1. Suppose we choose a = 0.7; then the forecast for week 3 would be
The actual observation for week 3 is 60; thus, the forecast for week 4 would be Week 4 Forecast = (1  0.7)(57.2) + (0.7)(60) = 59.16
Week 3 Forecast = (1  0.7)(88) + (0.7)(44) = 57.2
Because the simple exponential smoothing model requires only the previous forecast and the current timeseries value, it is very easy to calculate; thus, it is highly suitable for environments such as inventory systems, where many forecasts must be made. The smoothing constant a is usually chosen by experimentation in the same manner as choosing the number of periods to use in the moving average model. Different values of a affect how quickly the model responds to changes in the time series. For instance, a value of a = 0 would simply repeat last period’s forecast, whereas a = 1 would forecast the last period’s actual demand. The closer a is to 1, the quicker the model responds to changes in the time series, because it puts more weight on the actual current observation than on the forecast. Likewise, the closer a is to 0, the more weight is put on the prior forecast, so the model would respond to changes more slowly. EXAMPLE 9.9
Finding the Best Exponential Smoothing Model for Tablet Computer Sales
An Excel spreadsheet for evaluating exponential smoothing models for the Tablet Computer Sales data using values of a between 0.1 and 0.9 is shown in Figure 9.10. Note that in computing the error measures, the first row is not included
because we do not have a forecast for the first period, week 1. A smoothing constant of a = 0.6 provides the lowest error for all three metrics.
Excel has a Data Analysis tool for exponential smoothing. However, as opposed to the Moving Average tool, the chart generated by the Exponential Smoothing tool does correctly align the forecasts with the actual data. We will see this in the next example. EXAMPLE 9.10
Using Excel’s Exponential Smoothing Tool
In the Table Computer Sales example, from the Analysis group, select Data Analysis and then Exponential Smoothing. In the dialog (Figure 9.11), as in the Moving Average dialog, you must enter the Input Range of the timeseries data, the Damping Factor, (1  a)—not the smoothing constant as we have defined it—and the first cell of the Output
Range, which should be adjacent to the first data point. You also have options for labels, to chart output, and to obtain standard errors. The results are shown in Figure 9.12. You can see that the exponential smoothing model follows the pattern of the data quite closely, although it tends to lag with an increasing trend in the data.
▶▶ Figure
Chapter 9 Forecasting Techniques
9.10
Exponential Smoothing Forecasts for Tablet Computer Sales
▶▶ Figure
9.11
Exponential Smoothing Tool Dialog
▲▲ Figure
9.12
Excel Exponential Smoothing Forecasts for a = 0.6
365
366
Chapter 9 Forecasting Techniques
CHECK YOUR UNDERSTANDING 1. Explain how the simple moving average method works. 2. What are the three different types of error metrics that can be used to evaluate forecast accuracy? 3. Explain how simple exponential smoothing works. 4. Describe how to use the Excel Data Analysis tools for stationary time series.
Forecasting Models for Time Series with a Linear Trend For time series with a linear trend but no significant seasonal components, double exponential smoothing models or regressionbased models are more appropriate. These are based on the linear trend equation Ft + k = at + bt k (9.7)
That is, the forecast for k periods into the future from period t is a function of a base value at, also known as the level, and a trend, or slope, bt.
Double Exponential Smoothing In double exponential smoothing, the estimates of at and bt are obtained from the following equations:
at = aFt + 11  a21at  1 + bt  1 2 bt = b1at  at  1 2 + 11  b2bt  1
(9.8)
In essence, we are smoothing both parameters of the linear trend model. Initial values are chosen for a1 as A1 and b1 as A2  A1. Equations (9.8) must then be used to compute at and bt for the entire time series to be able to generate forecasts into the future. As with simple exponential smoothing, we are free to choose the values of a and b. The forecast for k periods beyond the last period (period T) is
EXAMPLE 9.11
FT + k = aT + bT (k) (9.9)
Double Exponential Smoothing
Figure 9.13 shows a portion of the Excel file Coal Production, which provides data on total tons produced from 1960 through 2011. The data appear to follow a linear trend. We will apply double exponential smoothing on just the first ten years of the data to illustrate the process. Figure 9.14 shows a spreadsheet implementation of this using a = 0.6 and b = 0.4 (Excel file Coal Production
Double Exponential Smoothing). We first initialize the values of a1 and b1: a1 = A1 = 434,329,000 b1 = A2  A1 = 420,423,000  434,329,000 = 13,906,000
Chapter 9 Forecasting Techniques
If we used equation (9.7) to forecast for period 2, we would obtain F2 = F1 + 1 = a1 + b1(1) = 434,329,000 + (13,906,000)(1) = 420,423,000 Note that this is the same as the actual value in period 2 because we used A2 to calculate b1, so we cannot make a true forecast until period 3. Continuing, we calculate a2 and b2 using formulas (9.8): a2 = aA2 + (1  a)(a1 + b1) = 0.6(420,423,000) + 0.4(434,329,000  13,906,000) = 420,423,000 b2 = b(a2  a1) + (1  b)b1 = 0.4(420,423,000  434,329,000) + 0.6(13,906,000) = 13,906,000 Then the forecast for period 3 is F3 = F2 + 1 = a2 + b2(1) = 420,423,000 + (13,906,000)(1) = 406,517,000
▶▶ Figure
9.13
Portion of Excel File Coal Production
▼▼ Figure
9.14
Excel Implementation of Double Exponential Smoothing
367
The other forecasts are calculated in a similar manner and are shown on the spreadsheet. The value of MAD for these forecasts is 23,471,063. We can experiment with different values of a and b to find a better fit. Using equation (9.9), the forecasting model for k periods beyond period 10 is F10 + k = a10 + b10(k) = 576,753,344 + 11,962,665(k) For example, the forecast for period 11 is F11 = 576,753,344 + 11,962,665(1) = 588,716,009, and the forecast for period 12 would be F12 = a10 + b10(2) = 576,753,344 + 11,962,665(2) = 600,678,674
368
Chapter 9 Forecasting Techniques
RegressionBased Forecasting for Time Series with a Linear Trend Equation 9.7 looks similar to the equation for simple linear regression. We introduced regression in the previous chapter as a means of developing relationships between a dependent and independent variables. Simple linear regression can be applied to forecasting using time as the independent variable.
EXAMPLE 9.12
Forecasting Using Trendlines
For the data in the Excel file Coal Production, a linear trendline, shown in Figure 9.15, gives an R2 value of 0.95 (the fitted model assumes that the years are numbered 1 through 52, not as actual dates). The model is Tons = 438,819,885.29 + 15,413,536.97 * Year
Thus, a forecast for 2012 (year 53) would be Tons = 438,819,885.29 + 15,413,536.97 * (53) = 1,255,737,345 Note, however, that the linear model does not adequately predict the recent drop in production after 2008.
In Chapter 8, we noted that an important assumption for using regression analysis is the lack of autocorrelation among the data. When autocorrelation is present, successive observations are correlated with one another; for example, large observations tend to follow other large observations, and small observations also tend to follow one another. This can often be seen by examining the residual plot when the data are ordered by time. Figure 9.16 shows the timeordered residual plot from the Excel Regression tool for the coal production example. The residuals do not appear to be random; rather, successive observations seem to be related to one another. This suggests autocorrelation, indicating that other approaches, called autoregressive models, are more appropriate. However, these are more advanced than the level of this book and are not discussed here.
▶▶ Figure
9.15
TrendlineBased Forecast for Coal Production Data
Chapter 9 Forecasting Techniques
▶▶ Figure
369
9.16
Residual Plot for Linear Regression Forecasting Model
CHECK YOUR UNDERSTANDING 1. What techniques are used to forecast time series with a linear trend? 2. Explain the process for double exponential smoothing. 3. How can trendlines be used in forecasting?
Forecasting Time Series with Seasonality Quite often, timeseries data exhibit seasonality. A “season” can be a year, a quarter, a month, or even a week, depending on the application. Most often, it is a year. We saw an example of this in Figure 9.2. When time series exhibit seasonality, different techniques provide better forecasts than the ones we have described.
RegressionBased Seasonal Forecasting Models One approach is to use linear regression. Multiple linear regression models with categorical variables can be used for time series with seasonality. To do this, we use dummy categorical variables for the seasonal components. EXAMPLE 9.13
RegressionBased Forecasting for Natural Gas Usage
With monthly data, as we have for natural gas usage in the Gas & Electric Excel file, we have a seasonal categorical variable with k = 12 levels. As discussed in Chapter 8, we construct the regression model using k  1 dummy variables. We will use January as the reference month; therefore, this variable does not appear in the model: Gas Usage = b0 + b1 Time + b2 February + b3 March + b4 April + b5 May + b6 June + b7 July + b8 August + b9 September + b10 October + b11 November + b12 December
This coding scheme results in the data matrix shown in Figure 9.17. This model picks up trends from the regression coefficient for time and seasonality from the dummy variables for each month. The forecast for the next January will be b0 + b1(25). The variable coefficients (betas) for each of the other 11 months will show the adjustment relative to January. For example, the forecast for next February will be b0 + b1(26) + b2(1), and so on. Figure 9.18 shows the results of using the Regression tool in Excel after eliminating insignificant variables (time
(continued )
370
Chapter 9 Forecasting Techniques
and Feb). Because the data show no clear linear trend, the variable time could not explain any significant variation in the data. The dummy variable for February was probably insignificant because the historical gas usage for both January and February were very close to each other. The R2 for this model is 0.971, which is very good. The final regression model is
▶▶ Figure
9.17
Data Matrix for Seasonal Regression Model
▼▼ Figure
9.18
Final Regression Model for Forecasting Gas Usage
Gas Usage = 
236.75  36.75 March  99.25 April 192.25 May  203.25 June  208.25 July 209.75 August  208.25 September 196.75 October  149.75 November 43.25 December
Chapter 9 Forecasting Techniques
371
HoltWinters Models for Forecasting Time Series with Seasonality and No Trend The methods we describe here and in the next section are based on the work of two researchers, C.C. Holt, who developed the basic approach, and P.R. Winters, who extended Holt’s work. Hence, these approaches are commonly referred to as Holt Winters models. HoltWinters models are similar to exponential smoothing models in that smoothing constants are used to smooth out variations in the level and seasonal patterns over time. For time series with seasonality but no trend, we can use one of two models, the HoltWinters additive seasonality model with no trend, Ft + k = at + St  s + k (9.10)
or the HoltWinters multiplicative seasonality model with no trend: Ft + k = atSt  s + k (9.11)
The additive model applies to time series with relatively stable seasonality, whereas the multiplicative model apples to time series whose amplitude increases or decreases over time. Therefore, a chart of the time series should be viewed first to identify the type of model to use. In both models, Sj is the seasonal factor for period j and s is the number of periods in a season. The forecast for period t + k is adjusted up or down from a level (at) by the seasonal factor. The multiplicative model is more appropriate when the seasonal factors are increasing or decreasing over time. HoltWinters Additive Seasonality Model with No Trend The level and seasonal factors are estimated in the additive model using the following equations:
Level component: at = a(At  St  s) + (1  a)at  1 Seasonal component: St = g(At  at) + (1  g)St  s
(9.12)
where a and g are smoothing constants. The first equation estimates the level for period t as a weighted average of the deseasonalized data for period t, (At  St  s), and the previous period’s level. The seasonal factors are updated as well using the second equation. The seasonal factor is a weighted average of the estimated seasonal component for period t, (At  at), and the seasonal factor for the last period of that season type. Then the forecast for the next period is Ft + 1 = at + St  s + 1. To begin, we need to estimate the level and seasonal factors for the first s periods (that is, the length of a season; for an annual season with quarterly data, this would be the first 4 periods; for monthly data, it would be the first 12 periods, and so on) before we can use the smoothing equations. We do this as follows:
at =
1 s Ai, for t = 1, 2, c, s (9.13) s ia =1
and
St = At  at, for t = 1, 2, c, s (9.14) Then we can use the smoothing equations to update at and St and calculate forecasts.
372
Chapter 9 Forecasting Techniques
EXAMPLE 9.14
Using the HoltWinters Additive Seasonality Model with No Trend
We will use the data for gas usage in the Excel file Gas & Electric. Looking at the chart in the Excel file (Figure 9.2), we see that the time series appears to have stable seasonality; thus the additive model would be most appropriate. We arbitrarily select a = 0.4 and g = 0.9.
S13 = g(A13  a13) + (1  g)S1 = 0.9(230  100.15) + (1  0.9)138.25 = 130.69 The forecast for period 14 is then
First, initialize the values of at and St for t = 1 to 12 using equations (9.13) and (9.14) as at = (244 + 228 + c + 88 + 199)/12 = 105.75, for t = 1 to 12. Then, S1 = A1  a1 = 244  105.75 = 138.25 S2 = A2  a2 = 228  105.75 = 122.25
F14 = a13 + S2 = 100.15 + 122.25 = 222.40 Other forecasted values are calculated in a similar manner. To forecast beyond the range of the last observed value, use formula (9.10) with t = T and k Ú 1: FT + k = aT + ST  s + k
and so on, up to S12. These are shown in the Excel file Holt Winters Additive Model for Seasonality (Figure 9.19). Using the seasonal factor for the first period, the forecast for period 13 is
The forecast for the period 25 (T = 24 and k = 1) is F25 = a24 + S13 = 105.80 + 130.69 = 236.49 To find F26, use T = 24 and k = 2:
F13 = a12 + S1 = 105.75 + 138.25 = 244
F26 = a24 + S14 = 105.80 + 134.45 = 240.25
We may now update the parameters based on the observed value for period 13, A13 = 230:
We can easily experiment with different values of the smoothing constants by changing the values of a and g in cells B3 and B4, and compare error measures such as MAD.
a13 = a(A13  S1) + (1  a)a12 = 0.4(230  138.25) + (1  0.4)105.75 = 100.15
HoltWinters Multiplicative Seasonality Model with No Trend The multiplicative seasonal model has the same basic smoothing structure as the additive seasonal model with some key differences:
Level component: at = a(At /St  s) + (1  a)at  1 Seasonal component: St = g(At /at) + (1  g)Sts
(9.15)
where a and g are again the smoothing constants. Here, At /St  s is the deseasonalized estimate for period t. Large values of a put more emphasis on this term in estimating the level for period t. The term At /at is an estimate of the seasonal factor for period t. Large values of g put more emphasis on this in the estimate of the seasonal factor. The forecast for the next period is Ft + 1 = atSt  s + 1. As with the additive model, first initialize the values for the level and seasonal factors. In this case, the formulas are
at =
1 s Ai, for t = 1, 2, c, s s ia =1
St = At >at, for t = 1, 2, c, s (9.16)
This model can be implemented on a spreadsheet in a similar fashion as the additive model.
Chapter 9 Forecasting Techniques
▲▲ Figure
373
9.19
Excel Implementation of HoltWinters Additive Model for Seasonality
HoltWinters Models for Forecasting Time Series with Seasonality and Trend Many time series exhibit both trend and seasonality. Such might be the case for growing sales of a seasonal product. These models combine elements of both the trend and seasonal models. Two types of HoltWinters smoothing models are often used. The HoltWinters additive seasonality model with trend is based on the equation
Ft + 1 = at + bt + St  s + 1 (9.17)
and the HoltWinters multiplicative seasonality model with trend is
Ft + 1 = 1at + bt 2St  s + 1 (9.18)
The additive model applies to time series with relatively stable seasonality and a trend, whereas the multiplicative model applies to time series whose amplitude increases or decreases over time along with a trend. As with the no trend case, a chart of the time series should be viewed first to identify the appropriate type of model to use. Three parameters, a, b, and g, are used to smooth the level, trend, and seasonal factors in the time series. HoltWinters Additive Seasonality Model with Trend This model is similar to the additive model that incorporates only seasonality, but with the addition of a trend component:
Level component: at = a(At  St  s) + (1  a)(at  1 + bt  1) Trend component: bt = b(at  at  1) + (1  b)bt  1 Seasonal component: St = g(At  at) + (1  g) St  s
(9.19)
374
Chapter 9 Forecasting Techniques
Here, a, b, and g are the smoothing parameters for level, trend, and seasonal components, respectively. The forecast for period t + 1 is Ft + 1 = at + bt + St  s + 1 (9.20)
The forecast for k periods beyond the last period of observed data (period T) is FT + k = aT + bT k + ST  s + k (9.21)
The initial values for level and seasonal factors are the same as in the additive seasonality model without trend, that is, formulas (9.13) and (9.14). The initial values for the trend component are
bt =
1 s (As+i  Ai) , for t = 1, 2, c, s s ia s =1
(9.22)
Basically, we are averaging As + 1  A1, As + 2  A2, c, and As + s  As. These become the trend factors for the first s periods. EXAMPLE 9.15
Using the HoltWinters Additive Model for Seasonality and Trend
We will use the data in the Excel file New Car Sales, which contains three years of monthly retail sales data. Figure 9.20 shows a chart of the data. Seasonality exists in the time series and appears to be stable, and there is an increasing trend; therefore, the HoltWinters additive model with seasonality and trend would appear to be the most appropriate model. We arbitrarily select a = 0.3, b = 0.2, and g = 0.9. Figure 9.21 shows the Excel implementation (Excel file Holt Winters Additive Model for Seasonality and Trend). The level and seasonal factors are initialized using formulas (9.13) and (9.14). The trend factors are initialized using formula (9.22): bt = [(42,227  39,810)>12 + (45,422  40,081)>12 + c + (49,278  44,186)>12]>12 = 462.48, for t = 1, 2, c, 12 Using Equation (9.20), the forecast for period 13 is F13 = a12 + b12 + S1 = 45,427.33 + 462.48 + (5,617.33) = 39,810.00
We may now update the parameters based on the observed value for period 13, A13 = 42,227: a13 = a(A13  S1) + (1  a)(a12 + b12) = 0.3(42,227  (5,617.33)) + 0.7(45,427.33 + 462.48) = 46,476.17 b13 = b(a13  a12) + (1  b)b12 = 0.2(46,476.17  45,427.33) + 0.8(462.48) = 579.75 S13 = g(A13  a13) + (1  g)S1 = 0.9(42,227  45427.33) + 0.1(5,617.33) = 4385.99 Other calculations are similar. To forecast period 37, we use formula (9.21): F36 + 1 = a36 + b36(1) + S36  12 + 1 = 49,267 units Notice from the chart that the additive model appears to fit the data extremely well. Again, we can easily experiment with different values of the smoothing constants to find a best fit using an error measure such as MAD or MSE.
HoltWinters Multiplicative Seasonality Model with Trend The HoltWinters multiplicative model is similar to the additive model for seasonality, but with a trend component:
Level component: at = a(At /St  s) + (1  a)(at  1 + bt  1) Trend component: bt = b(at  at  1) + (1  b)bt  1 Seasonal component: St = g(At /at) + (1  g) St  s
(9.23)
▶▶ Figure
Chapter 9 Forecasting Techniques
375
9.20
Portion of Excel File New Car Sales
▲▲ Figure
9.21
Excel Implementation of HoltWinters Additive Model for Seasonality and Trend
The forecast for period t + 1 is
Ft + 1 = (at + bt)St  s + 1 (9.24)
The forecast for k periods beyond the last period of observed data (period T) is
FT + k = (aT + bT k)ST  s + k (9.25)
376
Chapter 9 Forecasting Techniques
▼▼
TABLE 9.1
Forecasting Model Choice
No Seasonality
Seasonality
No trend
Simple moving average or simple exponential smoothing
HoltWinters additive or multiplicative seasonality models without trend or multiple regression
Trend
Double exponential smoothing
HoltWinters additive or multiplicative seasonality models with trend
Initialization is performed in the same way as for the multiplicative model without trend. Again, this model can be implemented in a similar fashion on a spreadsheet as the additive model.
Selecting Appropriate TimeSeriesBased Forecasting Models Table 9.1 summarizes the choice of forecasting approaches. CHECK YOUR UNDERSTANDING 1. Explain how to use dummy variables and set up worksheets to apply regression analysis for seasonal forecasting models. 2. Explain how to use the HoltWinters model for seasonality and no trend. 3. Explain how to use the HoltWinters model for seasonality and trend. 4. State what forecasting techniques are most appropriate for time series that may exhibit seasonality and/or trend.
Regression Forecasting with Causal Variables In many forecasting applications, other independent variables besides time, such as economic indexes or demographic factors, may influence the time series. For example, a manufacturer of hospital equipment might include such variables as hospital capital spending and changes in the proportion of people over the age of 65 in building models to forecast future sales. Explanatory/causal models, often called econometric models, seek to identify factors that explain statistically the patterns observed in the variable being forecast, usually with regression analysis. We will use a simple example of forecasting gasoline sales to illustrate econometric modeling.
EXAMPLE 9.16
Forecasting Gasoline Sales Using Simple Linear Regression
Figure 9.22 shows gasoline sales over ten weeks during June through August along with the average price per gallon and a chart of the gasoline sales time series with a fitted trendline (Excel file Gasoline Sales). During the summer months, it is not unusual to see an increase in sales as more people go on vacations. The chart
shows a linear trend, although R2 is not very high. The trendline is Sales = 4,790.1 + 812.99 Week Using this model, we would predict sales for week 11 as Sales = 4,790.1 + 812.99(11) = 13,733 Gallons
Chapter 9 Forecasting Techniques
▶▶ Figure
377
9.22
Gasoline Sales Data and Trendline
In the Gasoline Sales data, we also see that the average price per gallon changes each week, and this may influence consumer sales. Therefore, the sales trend might not simply be a factor of steadily increasing demand, but it might also be influenced by the average price per gallon. The average price per gallon can be considered as a causal variable. Multiple linear regression provides a technique for building forecasting models that incorporate not only time, but also other potential causal variables.
EXAMPLE 9.17
Incorporating Causal Variables in a RegressionBased Forecasting Model
For the Gasoline Sales data, we can incorporate the price/ gallon by using two independent variables. This results in the multiple regression model Sales = b0 + b1 Week + b2 Price>Gallon The results are shown in Figure 9.23, and the regression model is Sales = 72,333.08 + 508.67 * Week  16,463.20 * Price>Gallon
Notice that the R2 value is higher when both variables are included, explaining more than 86% of the variation in the data. If the company estimates that the average price for the next week will drop to $3.80, the model would forecast the sales for week 11 as Sales = 72,333.08 + 508.67 * 11  16,463.20 * 3.80 = 15,368 gallons
CHECK YOUR UNDERSTANDING 1. Explain the purpose of explanatory/causal (econometric) models. 2. Describe how to use multiple linear regression for causal forecasting.
The Practice of Forecasting Surveys of forecasting practices have shown that both judgmental and quantitative methods are used for forecasting sales of product lines or product families as well as for broad company and industry forecasts. Simple timeseries models are used for short and mediumrange forecasts, whereas regression analysis is the most popular method for longrange forecasting. However, many companies rely on judgmental methods far more than
378
Chapter 9 Forecasting Techniques
▶▶ Figure
9.23
Regression Results for Gasoline Sales
quantitative methods, and almost half judgmentally adjust quantitative forecasts. In this chapter, we focus on these three approaches to forecasting. In practice, managers use a variety of judgmental and quantitative forecasting techniques. Statistical methods alone cannot account for such factors as sales promotions, unusual environmental disturbances, new product introductions, large onetime orders, and so on. Many managers begin with a statistical forecast and adjust it to account for intangible factors. Others may develop independent judgmental and statistical forecasts and then combine them, either objectively by averaging or in a subjective manner. It is important to compare quantitatively generated forecasts to judgmental forecasts to see if the forecasting method is adding value in terms of an improved forecast. It is impossible to provide universal guidance as to which approaches are best, because they depend on a variety of factors, including the presence or absence of trends and seasonality, the number of data points available, the length of the forecast time horizon, and the experience and knowledge of the forecaster. Often, quantitative approaches will miss significant changes in the data, such as reversal of trends, whereas qualitative forecasts may catch them, particularly when using indicators as discussed earlier in this chapter.
ANALYTICS IN PRACTICE: Forecasting at NBCUniversal2 NBCUniversal (NBCU), a subsidiary of Comcast, is one of the world’s leading media and entertainment companies in the distribution, production, and marketing of entertainment, news, and information. The television broadcast year in the United States starts in the third week of September. The major broadcast networks announce their programming schedules for the new broadcast year in the middle of May. Shortly thereafter, the sale of 2
advertising time, which generates the majority of revenues, begins. The broadcast networks sell 60% to 80% of their airtime inventory during a brief period starting in late May and lasting two to three weeks. This sales period is known as the upfront market. Immediately after announcing their program schedules, the networks finalize their ratings forecasts and estimate the market demand. The ratings forecasts are projections of the numbers of people in each of several
Based on Srinivas Bollapragada, Salil Gupta, Brett Hurwitz, Paul Miles, and Rajesh Tyagi, “NBCUniversal Uses a Novel Qualitative Forecasting Technique to Predict Advertising Demand,” Interfaces, 38, 2 (March–April 2008): 103–111.
Chapter 9 Forecasting Techniques
demographic groups who are expected to watch each airing of the shows in the program schedule for the entire broadcast year. After they finalize their ratings projections and marketdemand estimates, the networks set the rate cards that contain the prices for commercials on all their shows and develop pricing strategies. Forecasting upfront market demand has always been a challenge. NBCU initially relied on historical patterns, expert knowledge, and intuition for estimating demand. Later, it tried timeseries forecasting models based on historical demand and leading economic indicator data and implemented the models in a Microsoft Excel–based system. However, these models proved to be unsatisfactory because of the unique nature of NBCU’s demand population. The timeseries models had fit and prediction errors in the range of 5% to 12% based on the historical data. These errors were considered reasonable, but the sales executives were reluctant to use the models because the models did not consider several qualitative factors that they believed influence the demand. As a result, they did not trust the forecasts that these models generated; therefore, they never used them. Analytics staff at NBCU then decided to develop a qualitative demand forecasting model that captures the knowledge of the sales experts. Their approach incorporates the Delphi method and “grassroots forecasting,” which is based on the concept of
379
Sean Pavone/Shutterstock
asking those who are close to the end consumer, such as salespeople, about the customers’ purchasing plans, along with historical data to develop forecasts. Since 2004, more than 200 sales and finance personnel at NBCU have been using the system to support sales decisions during the upfront market when NBCU signs advertising deals worth more than $4.5 billion. The system enables NBCU to sell and analyze pricing scenarios across all NBCU’s television properties with ease and sophistication while predicting demand with a high accuracy. NBCU’s sales leaders credit the system with having given them a unique competitive advantage.
CHECK YOUR UNDERSTANDING 1. Discuss practical considerations in forecasting. 2. How are the different types of forecasting methods used together? KEY TERMS Cyclical effect Delphi method Double exponential smoothing Econometric model Historical analogy HoltWinters additive seasonality model with no trend HoltWinters additive seasonality model with trend HoltWinters models HoltWinters multiplicative seasonality model with no trend HoltWinters multiplicative seasonality model with trend
Index Indicator Mean absolute deviation (MAD) Mean absolute percentage error (MAPE) Mean square error (MSE) Root mean square error (RMSE) Seasonal effect Simple exponential smoothing Simple moving average Smoothing constant Stationary time series Time series Trend
380
Chapter 9 Forecasting Techniques
CHAPTER 9 TECHNOLOGY HELP Excel Techniques Moving Average tool (Example 9.6): Select Data Analysis and then Moving Average from the Analysis group. In the dialog box, enter the Input Range of the data, the Interval (the value of k), and the first cell of the Output Range. To align the actual data with the forecasted values in the worksheet, select the first cell of the Output Range to be one row below the first value. We do not recommend using the chart or error options because the forecasts generated by this tool are not properly aligned with the data. Exponential Smoothing tool (Example 9.10): Select Data Analysis and then Exponential Smoothing from the Analysis group. In the dialog, enter
the Input Range of the timeseries data, the Damping Factor, which is (1  a)—not the smoothing constant a—and the first cell of the Output Range, which should be adjacent to the first data point. You also have options for labels, to chart output and to obtain standard errors. Analytic Solver Analytic Solver provides a set of tools for forecasting. See the online supplement Using Forecasting Techniques in Analytic Solver. We suggest that you first read the online supplement Getting Started with Analytic Solver Basic. This provides information for both instructors and students on how to register for and access Analytic Solver.
PROBLEMS AND EXERCISES Qualitative and Judgmental Forecasting 1. Identify some business applications in which judg
(that is, random, stationary, trend, seasonal, or cyclical).
mental forecasting techniques such as historical analogy and the Delphi method would be useful.
b. In forecasting the future, discuss whether all or
2. Search the Conference Board’s Web site to find
5. Use http://www.fao.org/faostat/en/#data to extract
business cycle indicators, and the components and methods adopted to compute the same. Write a short report about your findings.
an annual dataset for any topic of your interest that would aid in constructing a statistical forecasting model analysis. Download the relevant dataset into an Excel workbook, develop a line chart for the chosen variables, and forecast the data. What would be the forecasts for the next five years?
Statistical Forecasting Models 3. For each of the time series in the following Excel
files, construct a line chart of the data and identify the characteristics of the time series (that is, random, stationary, trend, seasonal, or cyclical). a. Closing Stock Prices b. Unemployment Rates c. New Car Sales d. Housing Starts e. Prime Rate 4. The Excel file Energy Production & Consumption
provides data on production, imports, exports, and consumption. a. Develop line charts for each variable and identify the characteristics of the time series
only a portion of the data should be used.
Forecasting Models for Stationary Time Series 6. For the data in the Excel file Airport Passengers, do
the following: a. Develop a spreadsheet for forecasting the number of airport passengers using a simple threeperiod moving average. b. Compute MAD, MSE, and MAPE error measures. 7. The Excel file New Registered Users provides data
on new registrations to an App on a monthly basis for four years. Compare the 3 and 12month moving average forecasts using the MAD criterion. Explain which model yields better results and why.
Chapter 9 Forecasting Techniques
381
8. The Excel file Closing Stock Prices provides data
15. Consider the data in the Excel file Internet Users.
for four stocks and the Dow Jones Industrial Average over a onemonth period. a. Develop a spreadsheet for forecasting each of the stock prices and the DJIA using a simple twoperiod and threeperiod moving average. b. Compute MAD, MSE, and MAPE and determine whether two or three moving average periods is better.
Use simple linear regression to forecast the data. What would be the forecast for the next three years?
9. The Excel file Stock Prices provides data for four
stocks and the stock exchange average over a 12day period. a. Use Excel’s Data Analysis Exponential Smoothing tool to forecast each of the stock prices using simple exponential smoothing with a smoothing constant of 0.3. b. Compute the MAD, MSE, and MAPE for each of
the models. c. Does a smoothing constant of 0.1 or 0.5 yield better results? 10. For the data in the Excel file Airport Passengers, do
the following: a. Develop a spreadsheet for forecasting the number of airport passengers using simple exponential smoothing with smoothing constants from 0.1 to 0.9 in increments of 0.1. b. Using MAD, MSE, and MAPE as guidance, find the best smoothing constant. Forecasting Models for Time Series with a Linear Trend 11. For the coal production example using double expo
Forecasting Time Series with Seasonality 16. Develop a multiple linear regression model with cat
egorical variables that incorporate seasonality for forecasting the deaths caused by accidents in the United States. Use the data for years 1976 and 1977 in the Excel file Accident Deaths. Use the model to generate forecasts for the next nine months, and compare the forecasts to actual observations noted in the data for the year 1978. 17. Develop a multiple regression model with categorical
variables that incorporate seasonality for forecasting sales using the last three years of data in the Excel file Auto Sales. 18. Develop a multiple regression model with categorical
variables that incorporate seasonality for forecasting 2010 housing starts using the 2008 and 2009 data in the Excel file Housing Starts. 19. Develop a multiple regression model with categorical
variables that incorporate seasonality for forecasting for the data in the Excel file Tourists Arrival. 20. For Example 9.14, experiment with the HoltWinters
Additive Model for Seasonality spreadsheet to find the best combination of a and g to minimize MAD. 21. Modify the Excel file HoltWinters Additive Model for
Seasonality (see Example 9.14) to implement the HoltWinters multiplicative seasonality model with no trend. 22. For Example 9.15, experiment with the HoltWinters
nential smoothing (Example 9.11), experiment with the spreadsheet model to find the best values of a and b that minimize MAD.
Additive Model for Seasonality and Trend spreadsheet model to find the best combination of a, b, and g to minimize MSE.
12. Consider the data in the Excel file Consumer Price
23. Modify the Excel file HoltWinters Additive Model
Index. Use the double exponential smoothing procedure to find forecasts for the next two years. 13. Consider the prices for the Stock Exchange in the
Excel file Stock Prices. Use simple linear regression to forecast the data. What would be the forecasts for the next three days? 14. Consider the data in the Excel file Consumer Price
Index. Use simple linear regression to forecast the data. What would be the forecasts for the next two years?
for Seasonality and Trend for the New Car Sales data (see Example 9.15) to implement the HoltWinters multiplicative seasonality model with trend. Try to find the best combination of a, b, and g to minimize MSE, and compare your results with Problem 22. 24. The Excel file CD Interest Rates provides annual aver
age interest rates on sixmonth certificate of deposits. Compare the HoltWinters additive and multiplicative models using a = 0.7, b = 0.3, g = 0.1, and a season of six years. Try to finetune the parameters using the MSE criterion.
382
Chapter 9 Forecasting Techniques
25. Using Table 9.1, determine the most appropriate fore
casting technique for the data in the Excel file DJIA December Close and implement the model. 26. Using Table 9.1, determine the most appropriate
forecasting technique for the data in the Excel file Mortgage Rates and implement the model. 27. Using Table 9.1, determine the most appropriate fore
casting technique for the data in the Excel file Prime Rate and implement the model. 28. Using Table 9.1, determine the most appropriate
forecasting technique for the data in the Excel file Treasury Yield Rates and implement the model. 29. Using Excel’s Moving Average tool for the data in
the Excel file Occupancy, identify the moving trend
for each of the hotels, motels, backpacker hostels, and holiday parks. Regression Forecasting with Causal Variables 30. Data in the Excel File Microchips shows the demand
for one type of chip used in industrial equipment from a small manufacturer. a. Construct a chart of the data. What appears to happen when a new chip is introduced? b. Develop a causal regression model to forecast demand that includes both time and the introduction of a new chip as explanatory variables. c. What would the forecast be for the next month if a new chip is introduced? What would it be if a new chip is not introduced?
CASE: PERFORMANCE LAWN EQUIPMENT An important part of planning manufacturing capacity is having a good forecast of sales. Elizabeth Burke is interested in forecasting sales of mowers and tractors in each marketing region as well as industry sales to assess future changes in market share. She also wants to forecast
future increases in production costs. Using the data in the Performance Lawn Equipment Database, develop forecasting models for these data and prepare a report of your results with appropriate charts and output from Excel.
CHAPTER
Introduction to Data Mining
10
Laborant/Shutterstock
LEARNING OBJECTIVES ■■
■■
■■
After studying this chapter, you will be able to:
Define data mining and some common approaches used in data mining. Explain how cluster analysis is used to explore and reduce data. Explain the purpose of classification methods and how to measure classification performance, and the use of training and validation data.
■■
■■
■■
Understand knearest neighbors and discriminant analysis for classification. Describe association rule mining and its use in market basket analysis. Use correlation analysis for causeandeffect modeling.
384
Chapter 10 Introduction to Data Mining
In an article in Analytics magazine, Talha Omer observed that using a cell phone to make a voice call leaves behind a significant amount of data. “The cell phone provider knows every person you called, how long you talked, what time you called and whether your call was successful or if was dropped. It also knows where you are, where you make most of your calls from, which promotion you are responding to, how many times you have bought before, and so on.”1 Considering the fact that the vast majority of people today use cell phones, a huge amount of data about consumer behavior is available. Similarly, many stores now use loyalty cards. At supermarkets, drugstores, retail stores, and other outlets, loyalty cards enable consumers to take advantage of sale prices available only to those who use the card. However, when they do, the cards leave behind a digital trail of data about purchasing patterns. How can a business exploit these data? If they can better understand patterns and hidden relationships in the data, they can not only understand buying habits but also customize advertisements, promotions, coupons, and so on for each individual customer and send targeted text messages and email offers (we’re not talking spam here, but registered users who opt into such messages). Data mining is a rapidly growing field of business analytics that is focused on better understanding characteristics and patterns among variables in large databases using a variety of statistical and analytical tools. Many of the tools that we have studied in previous chapters, such as data visualization, data summarization, PivotTables, and correlation and regression analysis, are used extensively in data mining. However, as the amount of data has grown exponentially, many other statistical and analytical methods have been developed to identify relationships among variables in large data sets and understand hidden patterns that they may contain. Many datamining procedures require advanced statistical knowledge to understand the underlying theory and special software to implement them. Therefore, our focus is on simple applications and understanding the purpose and application of datamining techniques rather than their theoretical underpinnings. In an optional online supplement, we describe the use of Analytic Solver for implementing datamining procedures.
The Scope of Data Mining Data mining can be considered part descriptive and part prescriptive analytics. In descriptive analytics, datamining tools help analysts to identify patterns in data. Excel charts and PivotTables, for example, are useful tools for describing patterns and analyzing data sets;
1
Talha Omer, “From Business Intelligence to Analytics,” Analytics (January/February 2011): 20. www.analyticsmagazine.org.
Chapter 10 Introduction to Data Mining
385
however, they require manual intervention. Regression analysis and forecasting models help us to predict relationships or future values of variables of interest. As some researchers observe, “the boundaries between prediction and description are not sharp (some of the predictive models can be descriptive, to the degree that they are understandable, and vice versa).”2 In most business applications, the purpose of descriptive analytics is to help managers predict the future or make better decisions that will impact future performance, so we can generally state that data mining is primarily a predictive analytic approach. Some common approaches in data mining include the following: ■■
■■
■■
■■
2
Cluster analysis. Some basic techniques in data mining involve data exploration and “data reduction”—that is, breaking down large sets of data into moremanageable groups or segments that provide better insight. We have seen numerous techniques earlier in this book for data exploration and data reduction. For example, charts, frequency distributions and histograms, and summary statistics provide basic information about the characteristics of data. PivotTables, in particular, are very useful in exploring data from different perspectives and for data reduction. Data mining software provide a variety of tools and techniques for data exploration that complement or extend the concepts and tools we have studied in previous chapters. This involves identifying groups in which the elements of the groups are in some way similar. This approach is often used to understand differences among customers and segment them into homogenous groups. For example, Macy’s department stores identified four types of customers defined by their lifestyle: “Katherine,” a traditional, classic dresser who doesn’t take a lot of risks and likes quality; “Julie,” neotraditional and slightly more edgy but still classic; “Erin,” a contemporary customer who loves newness and shops by brand; and “Alex,” the fashion customer who wants only the latest and greatest (they have male versions also).3 Such segmentation is useful in design and marketing activities to better target product offerings. These techniques have also been used to identify characteristics of successful employees and improve recruiting and hiring practices. Classification. Classification is the process of analyzing data to predict how to classify a new data element. An example of classification is spam filtering in an email client. By examining textual characteristics of a message (subject header, key words, and so on), the message is classified as junk or not. Classification methods can help predict whether a credit card transaction may be fraudulent, whether a loan applicant is high risk, or whether a consumer will respond to an advertisement. Association. Association is the process of analyzing databases to identify natural associations among variables and create rules for target marketing or buying recommendations. For example, Netflix uses association to understand what types of movies a customer likes and provides recommendations based on the data. Amazon.com also makes recommendations based on past purchases. Supermarket loyalty cards collect data on customers’ purchasing habits and print coupons at the point of purchase based on what was currently bought. Cause