Common method variance: From myth to procedural and statistical remedies. Exploring different methods to assess the risk of common method variance in survey-based research designs

The workshop will focus on the risks of common method variance (CMV) in survey-based behavioral research. The first part of the workshop (15 minutes) will be dedicated to understanding what CMV is and discussing debates around its mere existence (e.g., Richardson et al., 2009; Spector, 2006). In the second part of the workshop (15 minutes), regardless of the existing debates, we will discuss procedural remedies aimed at preventing the risks of CMV in survey-based research, such as the division of data collection by time, persons, and locations (e.g., Podsakoff et al., 2003). Procedural remedies have inherent strengths, weaknesses, and practical limitations. Third, we will address statistical remedies (60 minutes) for CMV, all post hoc analyses based on confirmatory factor analyses (CFA) to detect risks of CMV: the CFA marker technique and the unmeasured latent factor construct (ULMC) technique. For the ULMC technique, we will introduce the ConMET package for R (De Schutter, 2021), a package designed to conduct nested CFAs and contains the ULMC CMV test. Then, using an existing database, we will conduct tests using R scripts containing codes for both the ULMC and CFA marker techniques, to understand the pros and cons of each method. Because the ULMC tests will yield a positive result, we will explore additional tests to understand whether the threat is serious or not. The first and second parts of the workshop will be more theoretical, with slides presenting debates and procedural remedies, featuring practical examples in research. The second part of the workshop will also use some slides but will require participants to bring their laptops so that they can conduct the tests with us. The tests will require R and three packages: haven, lavaan, and ConMET. The database and scripts will be provided to participants before the workshop. Leander De Schutter will handle the technical and statistical discussions, while Corentin Hericher will lead the workshop and provide explanations about debates and remedies.