Common Beliefs and Reality About PLS

[Editor’s Note: We’re pleased to welcome Dr. Jörg Henseler, who was the corresponding author on the article, “Common Beliefs and Reality About PLS: Comments on Rönkkö and Evermann (2013)” from Organizational Research Methods.]

The extent to which an issue is raised by successive generations of researchers and practitioners is a subtle indicator of its importance. The benefits and limitations of partial least squares path modeling (PLS) is such an issue that has been heatedly debated across a wide variety of disciplines. Tying in with this stream of research, Rönkkö and Evermann (2013), in their recent Organizational Research Methods article, sought to examine “statistical myths and urban legends surrounding the often-stated capabilities of the PLS method and its current use in management and organizational research.” Based on a series of arguments and simulations studies, Rönkkö and Evermann (2013) conclude that “PLS results can be used to validate a measurement model is a myth” (p. 438); “the PLS path estimates cannot be used in NHST [null hypothesis significance testing]” (p. 439); “the small-sample-size capabilities of PLS are a myth” (p. 442); “PLS does not have [the capability to] reveal patterns in the data” (p. 442); “PLS lacks diagnostic tools” (p. 442); “PLS cannot be used to test models” (p. 442); and “PLS is not an appropriate choice for early-stage theory development and testing” (p. 442). In light of these results, the authors conclude that the use of PLS is difficult to justify and that researchers should rather revert to regression with summed scales or factor scores.

Considering the increasing popularity of PLS in the strategic management (Hair et al. 2012a), marketing (Hair et al. 2012b) and management information systems disciplines (Ringle et al. 2012; Figure 1), these claims are certainly alarming. But how is it possible that Rönkkö and Evermann (2013) cannot find even a single positive attribute of PLS which stands against the research of great minds such as the founder of PLS, Hermann Wold, and key contributor’s such as Jan-Bernd Lohmöller and Theo Dijkstra? Does the criticism really hold what Rönkkö and Evermann (2013) promise or do these authors create myths by chasing myths?

The Organizational Research Methods article “Common Beliefs and Reality about Partial Least Squares: Comments on Rönkkö & Evermann (2013),” authored by Jörg Henseler, Theo K. Dijkstra, Marko Sarstedt, Christian M. Ringle, Adamantios Diamantopoulos, Detmar W. Straub, Dave J. Ketchen, Joe F. Hair, G. Tomas M. Hult, and Roger Calantone provides answers to these questions and shows that none of the alleged shortcomings of PLS stands up. More precisely, we show that Rönkkö and Evermann’s (2013) surprising findings are not inherent in the PLS method but are rather the result of several limitations in their study, which indisputably limit the validity of the authors’ findings.

The major shortcoming of Rönkkö and Evermann’s (2013) study is that they neglect that PLS estimates a composite factor model, not a common factor model. Although the composite factor model is often a good approximation to the common factor model, there are important differences. Rönkkö and Evermann (2013) regard PLS simply as a suboptimal estimator of common factor models. But like a hammer is a suboptimal tool to fix screws, PLS is a suboptimal tool to estimate common factor models. In contrast, PLS is a useful tool for estimating composite factor models.

Another fundamental limitation of Rönkkö and Evermann’s (2013) study relates to their simulation design. Research on PLS has generated a multitude of different simulation studies that compare the technique’s performance with that of other approaches to structural equation modeling. These studies vary considerably in terms of their model set-ups. In this context and despite the fact that most recent simulation studies use quite complex models with a multitude of constructs and path relationships, Rönkkö and Evermann (2013) chose to use a two-construct model with a single path as their basis for their simulation. This, however, inevitably raises the question whether this model can indeed be considered representative of published research from an applied standpoint. Bearing this in mind, we revisited review studies on the use of PLS in strategic management, marketing, and information systems research. Out of the 532 PLS models being estimated in 306 journal articles, there was exactly one model (0.2 percent) with two constructs. More precisely, the average number of constructs was 7.94 in marketing, 7.50 in strategic management, and 8.12 in information systems, respectively. There are several other aspects of Rönkkö and Evermann’s (2013) simulation design which cast doubt on their findings, suggesting that their simulation model set-up is not remotely representative of research studies using PLS. Further limitations relate to implicit assumptions in their interpretation of the PLS method, over-stretched generalization of their findings, misinterpretation of the literature and reporting errors in their simulation results. By disclosing these shortcomings, our study re-establishes a constructive discussion of the PLS method and its properties.

On a more general level, our article should also be read as a reminder that there is no such thing as an estimation method that is best for every model, every distribution, every set of parameter values and every sample size. For all methods, no matter how impressive their pedigree (maximum likelihood being no exception), one can find situations where they do not work as advertised. One can always construct a setup where a given method, any method, ‘fails’. A (very) small sample or parameter values close to critical boundaries or distributions that are very skewed or thick-tailed etc., or any combination thereof will do the trick. It is just a matter of perseverance to find something that it is universally ‘wrong.’

A constructive attitude, one that aims to ascertain when PLS work well, how it can be improved would seem to be more conducive to improving the quality of research: “We believe that such debates are fruitful as long as they do not develop a ritualistic adherence to dogma and do not advocate one technique’s use as generally advantageous in all situations. Any extreme position that (often systematically) neglects the beneficial features of the other technique and may result in prejudiced boycott calls [citations removed], is not good research practice and does not help to truly advance our understanding of methods (or any other subject)” (Hair et al. 2012c, p. 313).

References
Hair, J. F., Sarstedt, M., Pieper, T. M., & Ringle, C. M. (2012a). Applications of partial least squares path modeling in management journals: a review of past practices and recommendations for future applications. Long Range Planning, 45(5-6), 320-340.
Hair, J. F., Sarstedt, M., Ringle, C. M., & Mena, J. A. (2012b). An assessment of the use of partial least squares structural equation modeling in marketing research. Journal of the Academy of Marketing Science, 40(3), 414-433.
Hair, J. F., Ringle, C. M., & Sarstedt, M. (2012c). Partial least squares: The better approach to structural equation modeling? Long Range Planning, 45(5-6), 312-319.
Henseler, J., Dijkstra, T. K., Sarstedt, M., Ringle, C. M., Diamantopoulos, A., Straub, D. W., Ketchen, D. J., Hair, J. F., Hult, G. T. M., & Calantone, R. J. (2014). Common beliefs and reality about partial least squares: Comments on Rönkkö & Evermann (2013). Organizational Research Methods, forthcoming.
Ringle, C. M., Sarstedt, M., & Straub, D. W. (2012). A critical look at the use of PLS-SEM in MIS Quarterly. MIS Quarterly, 36(1), iii-xiv.
Rönkko, M., & Evermann, J. (2013). A critical examination of common beliefs about partial least squares path modeling. Organizational Research Methods, 16(3), 425-448.

Click here to read the paper “Common Beliefs and Reality About PLS: Comments on Rönkkö and Evermann (2013)” from Organizational Research Methods. Want to know about all the latest from Organizational Research Methods? Click here to sign up for e-alerts!

Jörg Henseler, Institute for Management Research, Radboud University Nijmegen, Nijmegen, the Netherlands and ISEGI, Universidade Nova de Lisboa, Lisbon, Portugal

Theo K. Dijkstra, Faculty of Economics and Business, University of Groningen, Groningen, the Netherlands

Marko Sarstedt, Otto-von-Guericke University Magdeburg, Magdeburg, Germany and University of Newcastle, Callaghan, Australia

Christian M. Ringle, University of Newcastle, Callaghan, Australia and Hamburg University of Technology, Hamburg, Germany

Adamantios Diamantopoulos, University of Vienna, Vienna, Austria

Detmar W. Straub, J. Mack Robinson College of Business, Georgia State University, Atlanta, GA, USA

David J. Ketchen Jr., Raymond J. Harbert College of Business, Auburn University, Auburn, AL, USA

Joseph F. Hair, Coles College of Business, Kennesaw State University, Kennesaw, GA, USA

G. Tomas M. Hult, Broad College of Business, Michigan State University, East Lansing, MI, USA

Roger J. Calantone, Broad College of Business, Michigan State University, East Lansing, MI, USA