The first edition of our book was published just two years ago, in 2014. Why the need for a new edition when the first edition has been out only 2 years?
At the time we wrote the first edition, we were confident the interest in partial least squares structural equation modeling (PLSSEM) was increasing, but even we did not anticipate hot interest in the method would explode! Applications of PLS-SEM have grown exponentially in the past few years, and two journal articles we published before the first edition provide clear evidence of the popularity of PLS-SEM. The two articles have been the most widely cited in those journals since their publication—our 2012 article in the Journal of Academy of Marketing Science, “An Assessment of the Use of Partial Least Squares Structural Equation Modeling in Marketing Research,” cited more than 800 times according to Google Scholar, has been the number one highest impact article published in the top 20 marketing journals, according to Shugan’s list of most cited marketing articles (http://www.marketingscience.org; e.g., Volume 2, Issue 3). It has also been awarded the 2015 Emerald Citations of Excellence award. Moreover, our 2011 article in the Journal of Marketing Theory and Practice, “PLS-SEM: Indeed a Silver Bullet,” has surpassed more than 1,500 Google Scholar citations.
During this same timeframe, PLS-SEM has gained widespread interest among methods researchers as evidenced in a multitude of recent research papers that offer novel perspectives on the method. Very prominent examples include the rejoinders to Edward E. Rigdon’s (2012) Long Range Planning article by Bentler and Huang (2014); Dijkstra (2014); Sarstedt, Ringle, Henseler, and Hair (2014); and Rigdon (2014b) himself. Under the general theme “rethinking partial least squares path modeling,” this exchange of thoughts represents the point of departure of the most important PLS-SEM developments we expect to see in the next few years. Moreover, Rönkkö and Evermann’s (2013) paper in Organizational Research Methods offered an excellent opportunity to show how uninformed and blind criticism of the PLS-SEM method leads to misleading, incorrect, and false conclusions (see the rejoinder by Henseler et al., 2014). While this debate also nurtured some advances of PLS-SEM—such as the new heterotrait-monotrait (HTMT) criterion to assess discriminant validity (Henseler, Ringle, & Sarstedt, 2015)—we believe that it is important to reemphasize our previous call: “Any extreme position that (often systematically) neglects the beneficial features of the other technique and may result in prejudiced boycott calls, is not good research practice and does not help to truly advance our understanding of methods and any other research subject” (Hair, Ringle, & Sarstedt, 2012, p. 313).
Research has also brought forward methodological extensions of the original PLS-SEM method, for example, to uncover unobserved heterogeneity or to assess measurement model invariance. These developments have been accompanied by the release of SmartPLS 3, which implements many of these latest extensions in highly userfriendly software. This new release is much more than just a simple revision. It incorporates a broad range of new algorithms and major new features that previously had to be executed manually. For example, SmartPLS 3 runs on both Microsoft Windows and Mac OSX and includes the new consistent partial least squares algorithm, advanced bootstrapping features, the importance-performance map analysis, multigroup analysis options, confirmatory tetrad analysis to empirically assess the mode of measurement model, and additional segmentation techniques. Furthermore, new features augment data handling (e.g., use of weighted data) and the graphical user interface, which also includes many new options that support users running their analyses and documenting the results. In light of the developments in terms of PLS-SEM use, further enhancements, and extensions of the method and software support, a new edition of the book is clearly timely and warranted.
As noted in our first edition, the global explosion of data, often referred to as the “Age of Big Data,” is pushing the world toward data-driven discovery and decision making. The abundance of data presents both opportunities and challenges for scholars, industry, and government. While more data are available, there are not enough individuals with the analytical skills to probe and understand the data. Analysis requires a rigorous scientific approach dependent on knowledge of statistics, mathematics, measurement, logic, theory, experience, intuition, and many other variables affecting the situational context. Statistical analysis is perhaps the most important skill. While the other areas facilitate better understanding of data patterns, statistics provides additional substantiation in the knowledge- developing process. User-friendly software makes the application of statistics in the process efficient and cost-effective, in both time and money.
The increasing reliance on and acceptance of statistical analysis as well as the advent of powerful computer systems have facilitated the analysis of large amounts of data and created the opportunity for the application of more advanced next-generation analysis techniques. SEM is among the most useful advanced statistical analysis techniques that have emerged in the social sciences in recent decades. SEM is a class of multivariate techniques that combines aspects of factor analysis and regression, enabling the researcher to simultaneously examine relationships among measured variables and latent variables (assessment of measurement theory) as well as between latent variables (assessment of structural theory).
Considering the ever-increasing importance of understanding latent phenomena, such as consumer perceptions, expectations, attitudes, or intentions, and their influence on organizational performance measures (e.g., stock prices), it is not surprising that SEM has become one of the most prominent statistical analysis techniques today. While there are many approaches to conducting SEM, the most widely applied method since the late 1970s has been covariance-based SEM (CB-SEM). Since its introduction by Karl Jöreskog in 1973, CBSEM has received considerable interest among empirical researchers across virtually all social sciences disciplines. For many years, the predominance of LISREL, EQS, and AMOS, among the most wellknown software tools to perform this kind of analysis, led to a lack of awareness of the composite-based PLS-SEM approach as a very useful alternative approach to SEM. Originated in the 1960s by the econometrician Herman Wold (1966) and further developed in the years after (e.g., Wold, 1975, 1982, 1985), PLS-SEM has become an increasingly visible method in the social science disciplines.
Figure 1 summarizes the application of PLS-SEM in the top journals in the marketing and strategic management disciplines, as well as MIS Quarterly, the flagship journal in management information systems research. PLS-SEM use has increased exponentially in a variety of disciplines with the recognition that PLS-SEM’s distinctive methodological features make it an excellent alternative to the previously more popular CB-SEM approach. Specifically, PLS-SEM has several advantages over CB-SEM in many situations commonly encountered in social sciences research such as when sample sizes are small or when complex models with many indicators and model relationships are estimated. However, PLS-SEM should not be viewed simply as a less stringent alternative to CB-SEM but rather as a complementary modeling approach to SEM. If correctly applied, PLS-SEM indeed can be a silver bullet in many research situations.
Figure 1: Number of PLS-SEM Studies in Management, Marketing, and MIS Quarterly1 (Note: PLS-SEM studies published in MIS Quarterly were only considered from 1992 on).
[1For the selection of journals and details on the use of PLS-SEM in the three disciplines, see Hair, Sarstedt, Ringle, and Mena (2012); Hair, Sarstedt, Pieper, and Ringle (2012a); and Ringle, Sarstedt, and Straub (2012). Results for the most recent years have been added to the figure for the same selection of journals.]
PLS-SEM is evolving as a statistical modeling technique, and while there are several published articles on the method, until our first edition, there was no comprehensive book that explained the fundamental aspects of the method, particularly in a way that could be comprehended by individuals with limited statistical and mathematical training. This second edition of our book updates and extends the coverage of PLS-SEM for social sciences researchers and creates awareness of the most recent developments in a tool that will enable them to pursue research opportunities in new and different ways.
The approach of this book is based on the authors’ many years of conducting and teaching research, as well as the desire to communicate the fundamentals of the PLS-SEM method to a much broader audience. To accomplish this goal, we have limited the emphasis on equations, formulas, Greek symbols, and so forth that are typical of most books and articles. Instead, we explain in detail the basic fundamentals of PLS-SEM and provide rules of thumb that can be used as general guidelines for understanding and evaluating the results of applying the method. We also rely on a single software package (SmartPLS 3; http://www.smartpls.com) that can be used not only to complete the exercises in this book but also in the reader’s own research.
As a further effort to facilitate learning, we use a single case study throughout the book. The case is drawn from a published study on corporate reputation and we believe is general enough to be understood by many different areas of social science research, thus further facilitating comprehension of the method. Review and critical thinking questions are posed at the end of the chapters, and key terms are defined to better understand the concepts. Finally, suggested readings and extensive references are provided to enhance more advanced coverage of the topic.
We are excited to share with you the many new topics we have included in this edition. These include the following:
All examples in the edition are updated using the newest version of the most widely applied PLS-SEM software—SmartPLS 3. The book chapters and learning support supplements are organized around the learning outcomes shown at the beginning of each chapter. Moreover, instead of a single summary at the end of each chapter, we present a separate topical summary for each learning outcome. This approach makes the book more understandable and usable for both students and teachers. The Sage website for the book also includes other support materials to facilitate learning and applying the PLS-SEM method.
We would like to acknowledge the many insights and suggestions provided by the reviewers: Maxwell K. Hsu (University of Wisconsin), Toni M. Somers (Wayne State University), and Lea Witta (University of Central Florida), as well as a number of our colleagues and students. Most notably, we thank Jan-Michael Becker (University of Cologne), Adamantios Diamantopoulos (University of Vienna), Theo Dijkstra (University of Groningen), Markus Eberl (TNS Infratest), Anne Gottfried (University of Southern Mississippi), Verena Gruber (University of Vienna), Siegfried P. Gudergan (The University of Newcastle), Karl-Werner Hansmann (University of Hamburg), Jörg Henseler (University of Twente), Lucas Hopkins (Florida State University), Ida Rosnita Ismail (Universiti Kebangsaan Malaysia), Marcel Lichters (Harz University of Applied Sciences), David Ketchen (Auburn University), Gabriel Cepeda Carrión (University of Seville), José Luis Roldán (University of Seville), Lucy Matthews (Middle Tennessee State University), Roger Calantone (Michigan State University), Arthur Money (Henley Business School), Christian Nitzl (Universität der Bundeswehr München), Arun Rai (Georgia State University), Sascha Raithel (Freie Universität Berlin), Edward E. Rigdon (Georgia State University), Phillip Samouel (University of Kingston), Rainer Schlittgen (University of Hamburg), Manfred Schwaiger (Ludwig-Maxmillians University, Munich), Donna Smith (Ryerson University), Detmar W. Straub (Georgia State University), Sven Wende (SmartPLS GmbH), and Anita Whiting (Clayton State University) for their helpful remarks.
Also, we thank the team of doctoral student and research fellows at Hamburg University of Technology and Otto-von-Guericke-University Magdeburg—namely, Kathi Barth, Doreen Neubert, Sebastian Lehmann, Victor Schliwa, Katrin Engelke, Andreas Fischer, Nicole Richter, Jana Rosenbusch, Sandra Schubring, Kai Oliver Thiele, and Tabea Tressin—for their kind support. In addition, at SAGE we thank Vicki Knight, Leah Fargotstein, Yvonne McDuffee, and Kelly DeRosa. We hope this book will expand knowledge of the capabilities and benefits of PLS-SEM to a much broader group of researchers and practitioners. Last, if you have any remarks, suggestions, or ideas to improve this book, please get in touch with us. We appreciate any feedback on the book’s concept and contents!
Joe F. Hair, Jr., Kennesaw State University
G. Tomas M. Hult, Michigan State University
Christian M. Ringle, Hamburg University of Technology Germany, and the University of Newcastle, Australia
Marko Sarstedt, Otto-von-Guericke-University Magdeburg, Germany, and the University of Newcastle, Australia
Copyright © 2018 A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM)