Modeling Moderators in Psychological Networks

Abstract

It is an important goal for psychologists to develop and improve upon methods for describing multivariate relationships among observed variables. Psychological network models represent one class of methods for studying such relationships, and are being applied widely throughout psychological science. While these models have been shown to have a variety of diverse applications, they are limited by the fact that they currently only consider pairwise relationships among sets of variables. Specifically, they don’t take into consideration more complex relational structures such as those characterized by moderation effects and higher-order interactions. Moderation analysis, which focuses on these types of effects, is a common technique used within psychological research to help reveal the \textit{contexts} and \textit{conditions} under which different relationships may emerge or be observed. Thus, the goal of this research is to extend the psychological network framework to include moderator variables, as well as provide statistical tools and software to facilitate testing such models with psychological data.