On Common Method Variance: An Overview of Various Post-Hoc Methods to Detect and/or Control For Its Impact

Abstract

Common Method Variance (CMV) has been recognised as a threat to research validity by researchers in across many lines of inquiry. Often termed systematic measurement error, CMV has been demonstrated to alter (i.e. inflate or deflate) the true relationships among theoretical constructs in empirical studies (Bozionelos & Simmering, 2022; Chin et al., 2012; Podsakoff et al., 2003). As CMV arises from the instrumentation/data gathering process, many researchers choose to mitigate CMV pre hoc in the research design phase. Using multiple sources; separating the predictors and the criterion variables temporally, proximally, or psychologically; ensuring participants’ anonymity; and using a variety of measurement scales are some commonly used designs to minimize CMV. On the other hand, many researchers opt for post hoc statistical remedies to detect and ideally control for CMV in order to obtain accurate estimates of the true substantive effects being modelled. Some widely used statistical techniques include the Harman’s single-factor test, marker variable technique, the full collinearity approach, the unmeasured latent variable (UMLV) technique, and the directly measured latent variable technique. This talk will provide an overview and new findings on post-hoc approaches.

Biography

Wynne W. Chin is the C. T. Bauer Professor of Decision and Information Sciences in the C.T. Bauer College of Business at the University of Houston.

He is known for Partial Least Squares Path Modeling with his PLS-Graph software being the first graphical based PLS software developed in 1986 and released widely in 1990. Over his nearly 40 year span as an academic, he was the first to introduce the use of Monte Carlo simulation to the IS discipline for evaluating structural equation modeling algorithms. He also introduced bootstrapping for PLS analyses in 1988, two approaches for 2nd order models (1995), product indicator (1996, 2003) and Orthogonalising (2010) for interaction effects, PLS nonlinear modeling (1996), multigroup comparison via bootstrap with t-test formula (2000) and permutation (2003, 2016), bootstrap cross-validation (2005), latent marker variable for common method bias (2013), and bootstrap differential path tests (2013).

Wynne’s research has received over 106,000 citations, with a top-ten most cited paper in MIS Quarterly and top-five most cited papers in Information Systems Research, and ranked third overall in first authored articles published in MISQ and ISR for the period from 1990 through 2016, as well as a Google Scholar H-index of 71 placing him among the most impactful researchers in his field. Wynne is the 8th most cited researcher related to structural equation modeling. He is also ranked 9,994 in the world in the career database and 3,215 in the world in the 2019 single year database according to a 2019 research article list of the top 6.88 million scientists in the world for all disciplines in 22 fields and 176 subfields.

For his main field of Information and Communication Technologies, Wynne ranks 295 and 147 out of 570,025 scientists for career and single year contribution respectively. Wynne is a founding member of the Diffusion interest group in IT(DIGIT) in 1988 and the founding member and first President of the SigADIT (Special Interest Group in Adoption and Diffusion of IT). He was awarded a Fellow of the Association of Information Systems in 2013, garnered an AIS Distinguished Member - Cum Laude designation in 2020, and an AIS Technology Legacy Award (ATLAS) in 2021 recognizing those who have served the community in a significant way through their lifetime. Dr. Chin recently received the Farfel award (the highest faculty honor at UH). He currently serves as the treasurer for both the Texas Council of Faculty Senates and the National Council of Faculty Senates of which he is also a founding member.