Statistical Thinking

Apply statistics in business to achieve performance improvement

Statistical Thinking: Improving Business Performance, 3rd Edition helps managers understand the role of statistics in implementing business improvements. It guides professionals who are learning statistics in order to improve performance in business and industry. It also helps graduate and undergraduate students understand the strategic value of data and statistics in arriving at real business solutions. Instruction in the book is based on principles of effective learning, established by educational and behavioral research.

The authors cover both practical examples and underlying theory, both the big picture and necessary details. Readers gain a conceptual understanding and the ability to perform actionable analyses. They are introduced to data skills to improve business processes, including collecting the appropriate data, identifying existing data limitations, and analyzing data graphically. The authors also provide an in-depth look at JMP software, including its purpose, capabilities, and techniques for use.

Updates to this edition include:

• A new chapter on data, assessing data pedigree (quality), and acquisition tools
• Discussion of the relationship between statistical thinking and data science
• Explanation of the proper role and interpretation of p-values (understanding of the dangers of “p-hacking”)
• Differentiation between practical and statistical significance
• Introduction of the emerging discipline of statistical engineering
• Explanation of the proper role of subject matter theory in order to identify causal relationships
• A holistic framework for variation that includes outliers, in addition to systematic and random variation
• Revised chapters based on significant teaching experience
• Content enhancements based on student input

This book helps readers understand the role of statistics in business before they embark on learning statistical techniques.

Preface xiii

Introduction to JMP xvii

Part One Statistical Thinking Concepts 1

Chapter 1 Need for Business Improvement 3

Today’s Business Realities and the Need to Improve 4

We Now Have Two Jobs: A Model for Business Improvement 8

New Improvement Approaches Require Statistical Thinking 12

Principles of Statistical Thinking 17

Applications of Statistical Thinking 22

Summary and Looking Forward 23

Exercises: Chapter 1 24

Notes 25

Chapter 2 Data: The Missing Link 27

Why Do We Need Data? 28

Types of Data 29

All Data are Not Created Equal 32

Practical Sampling Tips to Ensure Data Quality 34

What about Data Quantity? 38

Every Data Set Has a Story: The Data Pedigree 40

The Measurement System 42

Summarizing Data 48

Summary and Looking Forward 52

Exercises: Chapter 2 52

Notes 54

Chapter 3 Statistical Thinking Strategy 55

Case Study: The Effect of Advertising on Sales 56

Case Study: Improvement of a Soccer Team’s Performance 62

Statistical Thinking Strategy 71

Variation in Business Processes 76

Synergy between Data and Subject Matter Knowledge 82

Dynamic Nature of Business Processes 84

Value of Graphics—Discovering the Unexpected 86

Summary and Looking Forward 89

Project Update 89

Exercises: Chapter 3 90

Notes 91

Chapter 4 Understanding Business Processes 93

Examples of Business Processes 94

SIPOC Model for Processes 100

Identifying Business Processes 102

Analysis of Business Processes 103

Systems of Processes 119

Summary and Looking Forward 122

Project Update 123

Exercises: Chapter 4 124

Notes 126

Part Two Holistic Improvement: Frameworks and Basic Tools 127

Chapter 5 Holistic Improvement: Tactics to Deploy Statistical Thinking 129

Case Study: Resolving Customer Complaints of Baby Wipe Flushability 130

The Problem-Solving Framework 137

Case Study: Reducing Resin Output Variation 141

The Process Improvement Framework 147

Statistical Engineering 153

Statistical Engineering Case Study: Predicting Corporate Defaults 154

A Framework for Statistical Engineering Projects 158

Summary and Looking Forward 164

Project Update 165

Exercises: Chapter 5 166

Notes 167

Chapter 6 Process Improvement and Problem-Solving Tools 169

Practical Tools 172

Knowledge-Based Tools 191

Graphical Tools 207

Analytical Tools 228

Summary and Looking Forward 265

Project Update 265

Exercises: Chapter 6 266

Notes 271

Part Three Formal Statistical Methods 273

Chapter 7 Building and Using Models 275

Examples of Business Models 276

Types and Uses of Models 279

Regression Modeling Process 282

Building Models with One Predictor Variable 290

Building Models with Several Predictor Variables 307

Multicollinearity: Another Model Check 315

Some Limitations of Using Observational Data 317

Summary and Looking Forward 319

Project Update 321

Exercises: Chapter 7 321

Notes 346

Chapter 8 Using Process Experimentation to Build Models 347

Randomized versus Observational Studies 348

Why Do We Need a Statistical Approach? 350

Examples of Process Experiments 355

Problem-Solving and Process Improvement are Sequential 364

Statistical Approach to Experimentation 365

Two-Factor Experiments: A Case Study 372

Three-Factor Experiments: A Case Study 378

Larger Experiments 385

Blocking, Randomization, and Center Points 387

Summary and Looking Forward 389

Project Update 391

Exercises: Chapter 8 391

Notes 399

Chapter 9 Applications of Statistical Inference Tools 401

Examples of Statistical Inference Tools 404

Process of Applying Statistical Inference 408

Statistical Confidence and Prediction Intervals 412

Statistical Hypothesis Tests 424

Tests for Continuous Data 435

Test for Discrete Data: Comparing Two or More Proportions 441

Test for Regression Analysis: Test on a Regression Coefficient 442

Sample Size Formulas 443

Summary and Looking Forward 448

Project Update 449

Exercises: Chapter 9 450

Notes 454

Chapter 10 Underlying Theory of Statistical Inference 455

Applications of the Theory 456

Theoretical Framework of Statistical Inference 458

Probability Distributions 463

Sampling Distributions 479

Linear Combinations 486

Transformations 490

Summary and Looking Forward 510

Project Update 511

Exercises: Chapter 10 511

Notes 514

Appendix A Effective Teamwork 515

Appendix B Presentations and Report Writing 525

Appendix C More on Surveys 531

Appendix D More on Regression 539

Appendix E More on Design of Experiments 553

Appendix F More on Inference Tools 567

Appendix G More on Probability Distributions 571

Appendix H DMAIC Process Improvement Framework 577

Appendix I t Critical Values 587

Appendix J Standard Normal Probabilities (Cumulative z Curve Areas) 589

Index 593

DR. ROGER W. HOERL is an associate professor at Union College where he teaches statistics, engineering statistics, design of experiments, regression analysis, and big data analytics. Previously, he led the Applied Statistics Laboratory at GE Global Research.

DR. RONALD D. SNEE is founder and president of Snee Associates, an authority on designing and implementing organizational improvement and cost-reduction solutions. Prior to this role, he worked at the DuPont Company in a variety of assignments. Snee has co-authored five books and published more than 330 articles on process improvement, quality, and statistics.

Data and statistical analysis are increasingly integral parts of modern business improvement, including the lean six sigma and holistic improvement approaches. No longer the purview of high-level specialists, statistics is now a required competency for managers and business students, because performance improvement is a day-to-day must for every business.

Because statistical analysis has been infused into the contemporary business context, we need to understand business performance in a way that incorporates the role of statistics. Data and analysis are no longer the cherries on top—they are some of the main ingredients in today's top performing organizations. Uniquely, Statistical Thinking: Improving Business Performance recognizes the need for an integrated approach and gives readers a real-world perspective that goes far beyond the mathematical formulas.

This Third Edition features a long list of updates and revisions that further align the content with current business techniques and educational best practices. The authors have walked the walk, using their own statistical analysis of student feedback to identify major improvements to the text. Additional topics, more visuals, clearer descriptions, and a new chapter focusing on data quality are just a few of the improvements.

Authors and business performance gurus Roger Hoerl and Ron Snee have received acclaim for their engaging, reality-first approach to statistics education. Learners gain a top-down view of statistics and what we are really trying to accomplish when we analyze business data. Each analysis topic is first presented in context, showing how it works in practice before teasing apart the theoretical details. The result is a thoroughly modern book that will help readers develop, not just the skillset, but the mindset to improve business performance using statistics.

A STATE-OF-THE-ART INTRODUCTION TO BUSINESS STATISTICS

In today's business environment, performance improvement must be a constant. The most effective managers and performance specialists are those who can see statistics everywhere they look. They understand the "big picture" of business processes and can apply data analysis methodologies with purpose, deriving actionable conclusions from data and helping organizations toward their goals. Statistical Thinking is the only business statistics book that will help you develop these big picture skills while training you in the use of important tools, including JMP software and key methodologies.

New to this Third Edition:

• A renewed focus on real-world issues of current importance in business, including data science, data visualization, p values and statistical significance, and statistical engineering
• An all-new chapter on data, data pedigree (quality), tools for acquiring data, and evaluating data measurement systems
• A practical framework for understanding outliers, systematic variation, and random variation, enabling you to understand and explain anomalies in any dataset
• Thorough content revisions, including updated graphs and tables, based on firsthand teaching experience and student input