Explaining Change: Theorizing and Testing Dynamic Mediated Longitudinal Relationships

Abstract

Many disciplines of scholarship have developed theories that involve dynamic mediated (and multilevel) relationships among constructs. However, most research does not hypothesize or test these dynamic relationships in a manner consistent with theory. In this article, the authors address this disconnect by first noting the theoretical and methodological limitations of ignoring dynamic mediated (and multilevel) relationships. Specifically, the authors show that theory testing suffers and statistical conclusions are often erroneous when dynamic mediation is ignored. The authors then present several ways of conceptualizing dynamic mediated relationships and then turn to summarizing two statistical models for analyzing such data. They conclude with a brief example from a team performance context.