Overview and Updates
Table of Contents
Updates and Corrections
Now in the eighth edition, Multivariate Data Analysis has changed considerably since its first publication in 1979. Throughout this time, the authors have strived to serve both academicians and applied researchers with a text that presents both the basic techniques as well as the “cutting-edge” developments” in a presentation style amenable to both audiences: an applications-oriented introduction to multivariate analysis for the nonstatistician. We encourage you to see what is new or been revised in the eighth edition as well as periodically check back with to view updates to the chapters as well as any corrections that should be noted.
Overview of the Eighth Edition
The past decade has seen an explosion in the interest and application of data analytics in both academic research and decision-making in all types of organizations. The emergence of Big Data has provided a newfound wealth of information available to address questions in all fields of study. To meet that need, researchers have continued their development of the traditional techniques as well as exploring new avenues of analysis. The eighth edition of Multivariate Data Analysis provides an updated perspective on data analysis of all types of data as well as introducing some new perspectives and techniques that are foundational in today’s world of analytics.
Among the specific additions and updates in the eight edition:
- PLS_SEM — New chapter on partial least squares structural equation modeling (PLS-SEM), an emerging technique with equal applicability for researchers in the academic and organizational domains.
- Integration of the implications of Big Data into each of the chapters, providing some understanding of the role of multivariate data analysis in this new era of analytics
- Causal Inference and Multi-level/Panel Models — Extended discussions of emerging topics, including causal treatments/inference (i.e., causal analysis of non-experimental data as well as discussion of propensity score models) along with multi-level and panel data models (extending regression into new research areas and providing a framework for cross-sectional/time-series analysis).
- Updates — Updates in each chapter of technical improvements (e.g., multiple imputation for missing data treatments) as well as the merging of basic principles from the fields of data mining and its applications.
- Additional SEM Topics — In addition to the new PLS-SEM chapter, the chapters on SEM have greater emphasis on psychometrics and scale development, discussions on the use of reflective versus formative scaling, describes an alternative approach for handing interactions (orthogonal moderators), higher order models, multi-group analyses, Bayesian SEM, and updated availability software availability (e.g., Lavaan and SmartPLS). The multi-group discussion also includes an alternative to partial metric invariance when cross-group variance problems are small.
- Online resources — For researchers including continued coverage from past editions of all of the analyses from the latest versions of both SAS and SPSS (commands and outputs)
Overview of the Seventh Edition
More than 30 years ago when the first edition of Multivariate Data Analysis was published, we could not have imagined the applications of multivariate statistics would be as pervasive as they are today. This continued interest has contributed to the acceptance of the past six editions of this text and the demand for this 7th edition. In approaching this revision, we continually strive to reduce our reliance on statistical notation and terminology and instead to identify the fundamental concepts which affect application of these techniques and then express them in simple terms—the result being an applications-oriented introduction to multivariate analysis for the non-statistician.
One new feature of the seventh edition is the development of separate versions for the U.S. domestic market and the international community. While many of the concepts are similar, each edition has been tailored to match the needs of the different user. Of particular note is the addition of a chapter on canonical correlation to the international edition due to increased demand for coverage of that topic.
Based on much positive feedback, the “Rules of Thumb” for the application and interpretation of the various techniques have been expanded in this edition, including important issues like sample size. The rules of thumb are highlighted throughout the chapters to facilitate their use. We are confident these guidelines will facilitate your utilization of the techniques.
Special thanks are due to Pei-ju Lucy Ting and Hsin-Ju Stephanie Tsai, both from University of Manchester, for the revision of the chapter on canonical correlation analysis. They updated this chapter with an example using the HBAT database, added recently published material, and reorganized it to facilitate understanding.
What's New in the Domestic Edition
A primary objective of the 7th edition was to streamline coverage of the materials, with particular emphasis on making each chapter shorter and simpler in its organization, with chapters typically focusing on a single topic. For example, multiple discriminant analysis and logistic regression are separate chapters, as are multidimensional scaling and correspondence analysis. Two chapters, cluster analysis and conjoint, were extensively revised to more effectively demonstrate straightforward approaches to obtain solutions.
We have also undertaken a substantial expansion and reorganization in coverage of structural equations modeling. We now have four chapters on this increasingly important technique plus a comparison with partial least squares. Chapter 12 provides an overview of structural equation modeling; Chapter 13 focuses on confirmatory factor analysis; Chapter 14 covers issues in estimating and testing structural models; and Chapter 15 reviews a few more advanced topics in both confirmatory factor analysis and structural equations modeling, such as testing higher-order factor models, group models, and moderating and mediating variables. These four chapters provide a comprehensive overview and explanation of this technique.
What's New in the International Edition
As noted earlier, special thanks are due to Pei-ju Lucy Ting and Hsin-Ju Stephanie Tsai, both from University of Manchester, for a new chapter on canonical correlation analysis (Chapter 5). They updated this chapter with an example using the HBAT database, added recently published material, and reorganized it to facilitate understanding.
This edition is shorter and simpler in its organization, with all chapters revised to incorporate advances in technology, and several chapters have undergone more extensive change. Two chapters, cluster analysis and conjoint, were extensively revised to more effectively demonstrate straightforward approaches to obtain solutions.
Another major change is the expansion and reorganization in coverage of structural equations modeling. Chapter 11 provides an overview of structural equation modeling. Chapter 12 then focuses on confirmatory factor analysis, issues in estimating and testing structural models, and advanced topics in both confirmatory factor analysis and structural equations modeling, including PLS. As always, we strived to eliminate and minimize the use of technical terms and mathematical and statistical notation that often is confusing.
Overview of the Sixth Edition
In approaching this revision, we have tried to embrace both academic and applied researchers with a presentation strongly grounded in statistical techniques, but focusing on design, estimation, and interpretation. We continually strive to reduce our reliance on statistical notation and terminology and instead to identify the fundamental concepts which affect our use of these techniques and then express them in simple terms. Our commitment remains to provide a firm understanding of the statistical and managerial principles underlying multivariate analysis so as to develop a “comfort zone” not only for the statistical but also the practical issues involved.
What's New in the Sixth Edition
- The most obvious change in the sixth edition is the new database—HBAT. The emphasis on improved measurement, particularly multi-item constructs, led to its development. After substantial testing we believe it provides an expanded teaching tool with various techniques that are comparable to the HATCO database, which will still be available on the book’s Web site.
- A second major addition is “Rules of Thumb” for the application and interpretation of the various techniques. The rules of thumb are highlighted throughout the chapters to facilitate their use. We are confident these guidelines will facilitate your utilization of the techniques.
- A third major change to the text is a substantial expansion in coverage of structural equations modeling. We now have three chapters on this increasingly important technique. Chapter 10 provides an overview of structural equation modeling, Chapter 11 focuses on confirmatory factor analysis, and Chapter 12 covers issues in estimating structural models. These three chapters provide a comprehensive introduction to this technique.
- Finally, there is an expanded set of supplementary materials for instructors. PowerPoint slides are available for each chapter, along with an updated Instructor’s Manual. All of these materials and more are available in the “Adopter’s Section” of this website.
What's Expanded and Updated
Each chapter has been revised to incorporate advances in technology, and several chapters have undergone more extensive change.
- Chapter 5, “Multiple Discriminant Analysis and Logistic Regression,” provides complete coverage of analysis of categorical dependent variables, including both discriminant analysis and logistic regression. Particular emphasis was placed on an expanded discussion of logistic regression which includes an illustrative example using the HBAT database.
- Chapter 7, “Conjoint Analysis,” contains a revised and updated examination of “cutting edge” issues relating to both research design and model estimation. This technique is rapidly evolving as a critical tool in today’s academic and practitioner research efforts and our focus is on both practical and conceptual issues that impact use of this technique.
- Finally, Chapter 10, “Structural Equation Modeling,” has been updated and expanded to three chapters in order to reflect the many changes in the field in the past several years. Each chapter also contains a number of expanded topics and example analyses such as testing higher-order factor models, group models, and moderating and mediating variables.
This edition is shorter and simpler in its organization, with all chapters revised to incorporate advances in technology, and several chapters have undergone more extensive change. Two chapters, cluster analysis and conjoint, were extensively revised to more effectively demonstrate straightforward approaches to obtain solutions.
Another major change is the expansion and reorganization in coverage of structural equations modeling. Chapter 11 provides an overview of structural equation modeling. Chapter 12 then focuses on confirmatory factor analysis, issues in estimating and testing structural models, and advanced topics in both confirmatory factor analysis and structural equations modeling, including PLS. As always, we strived to eliminate and minimize the use of technical terms and mathematical and statistical notation that often is confusing.
Updates and Corrections
With every new edition there is the need to provide updates on new topics as well as correct any errata identified in the chapters. We will always correct errata in the next possible printing, but for those users who may encounter an issue in an earlier edition, we will list corrections here as well.
Eighth Edition
- None at this time.
Seventh Edition
- None at this time.
Sixth Edition
Chapter 2 – Examining Your Data
- Corrected Figure 2-4: Two small changes are noted in the corrected figure
Chapter 8 – Cluster Analysis
- Exact replication of the results will require some dataset re-ordering as described in this documentation.