This thesis investigates how current marketing research deals with simple moderation analysis in marketing applications of partial least squares structural equation modeling (PLS-SEM). In this context, different approaches to generate the interaction term as well as options of data treatment before analysis, among others, pose challenges to researchers. The goal of this thesis is therefore to highlight the most common sources of error through a review of 33 relevant empirical articles from the PLS-SEM context and to derive recommendations for research. Using a framework that shows how to preferably generate the interaction term, it is found that about 70% of the analyzed studies use a suboptimal approach. Furthermore, the product-indicator approach is the most frequently used but only in exceptional cases the optimal approach. Additionally, both the approach used as well as the effect sizes of moderation effects are generally rarely specified and finally uncertainties in the interpretation of effects in moderation analyses can be observed. This confirms the urgency of the need for educating researchers in terms of moderation analysis in the context of PLS-SEM to minimize common mistakes in the future that distort results of moderation analyses or make their meaning incorrect.
Keywords: Moderation analysis; interaction effect; moderator; PLS-SEM; Partial least squares structural equation modeling.