Abstract Consumer researchers often compare a proposed process across contexts (i.e., moderated mediation) or across mediators (e.g., ruling out an alternative process in parallel mediation). This paper aims to help researchers in mapping process-related theoretical hypotheses onto statistical coefficients and in reporting their results. Researchers can formulate a variety of hypotheses about the conditional indirect effect (CIE) of a predictor X on an outcome Y through a mediator M, such as: (A) the CIE is greater for moderator condition W0 than W1, (B) the CIE is only positive for W0, (C) the CIE is positive for W0 and negative for W1. For example, if a researcher’s hypotheses align with case A, they must test and report the difference between the two conditional indirect effects (i.e., the index of moderated mediation). Reporting that the indirect effect is significant for W0 and non-significant for W1 would be insufficient for case A, but appropriate for case B. We generalize these examples in two tutorials–for moderated mediation and for parallel mediation–to help researchers 1) connect theory to testable predictions, 2) select the appropriate statistical model, and 3) report results transparently and consistently. We provide concrete examples of pre-registrations, data analyses, and manuscript reports.
Cadario et al. (Thu,) studied this question.