Time-Varying Effect Modeling to Address Novel Questions in Developmental Research

Time-varying effect modeling (TVEM) is a statistical approach that enables researchers to estimate associations between variables as continuous functions of time, thus holding potential to address new questions in developmental science. First, we present a conceptual introduction to unique developmental questions that TVEM is well suited to address, such as questions about changes across development in the strength of associations between risk factors and problematic outcomes. We then summarize the key contributions of TVEM for developmental research. Second, we summarize technical details that are essential for applying TVEM to empirical research. Third, we present an empirical demonstration of how to apply TVEM to address developmental research questions. In this demonstration, we use TVEM to examine whether the strength of association between marijuana use and depressive symptoms varies across high school ages and identify specific ages at which this association differs for male and female adolescents. The syntax used to estimate these models and create figures showing relevant coefficient functions is included in an appendix. We conclude this chapter with a discussion of the demonstration, the application of TVEM to other time metrics (for example, to examine within-day processes by modeling intensive longitudinal data), and future directions for the application and development of TVEM in developmental research.