A critical look at the use of PLS-SEM in MIS Quarterly

Wold’s (1974; 1982) partial least squares structural equation modeling (PLS-SEM) ap-proach and the advanced PLS-SEM algorithms by Lohmoller (Lohmoller 1989) have enjoyed steady popularity as a key multivariate analysis methods in management infor-mation systems (MIS) research (Gefen et al. 2011). Chin’s (1998b) scholarly work and technology acceptance model (TAM) applications (e.g., Gefen and Straub 1997) are milestones that helped to reify PLS-SEM in MIS research. In light of the proliferation of SEM techniques, Gefen et al. (2011), updating Gefen et al. (2000), presented a compre-hensive, organized, and contemporary summary of the minimum reporting requirements for SEM applications. Such guidelines are of crucial importance for advancing research for several reasons. First, researchers wishing to apply findings from prior studies or wanting to contribute to original research must comprehend other researchers’ decisions in order to under-stand the robustness of their findings. Likewise, when studies arrive at significantly different results, the natural course is to attempt explaining the differences in terms of the theory or concept employed, the empirical data used, and how the research method was applied. A lack of clarity on these issues, including the methodological applications, contradicts the goals of such studies (Jackson et al. 2009). Even worse, the misapplication of a technique may result in misinterpretations of empirical outcomes and, hence, false conclusions. Against this background, rigorous research has a long-standing tradition of critically reviewing prior practices of reporting standards and research method use (e.g., Boudreau et al. 2001). While the use of covariance-based SEM (CB-SEM) techniques has been well documented across disciplines (e.g., Medsker et al. 1994; Shook et al. 2004; Steenkamp and Baumgartner 2000), few reviews to date have investigated usage practices specific to PLS-SEM (see, however, Gefen et al. 2000). Previous reviews of such research practices were restricted to strategic management (Hulland 1999) and, more recently, marketing (Hair et al. 2012; Henseler et al. 2009), and accounting (Lee et al. 2011). The question arises as to how authors publishing in top IS journals such as MIS Quarterly have used PLS-SEM thus far, given the SEM recommendations of Gefen et al. (2011). By relating Gefen et al.’s (2011) reporting guidelines to actual practice, we attempt to identify potential problematic areas in PLS-SEM use, problems which may explain some of the criticism of how it has been applied (e.g., Marcoulides et al. 2009; Marcoulides and Saunders 2006). By reviewing previous PLS-SEM research in MIS Quarterly, we can hopefully increase awareness of established reporting standards. The results allow researchers to further improve the already good reporting practices that have been established in MIS Quarterly and other top journals and thus could become blueprints for conducting PLS-SEM analysis in other disciplines such as strategic management and marketing.