ABSTRACT
This chapter describes how to incorporate moderators in mediation analysis. The strength and form of mediation effects may depend on other variables, called moderators. A general approach to modeling mediation and moderation is described with a moderator with two different values. Moderator effects have been studied more extensively studied than mediator effects both in application and as a methodology, most notably in the context of analysis of variance. From a substantive perspective, interactions are especially interesting as they imply that an observed relation between an independent variable and a dependent variable can be strengthened, weakened, removed, or made opposite in sign when the third variable is considered. The test of the XM interaction has important substantive implications as well as providing a test of an assumption of the single mediator model. One example of a moderation and mediation model that explicitly includes a baseline score is the mediated baseline by treatment interaction model.
