Further Applications and Future Directions

In this final chapter, we begin with a discussion of several considerations related to the interpretation of results from time-varying effect modeling (TVEM), including time lags and conceptual differences between TVEM and multilevel modeling for analyzing longitudinal data. We then present several ways that TVEM holds potential to advance the social, behavioral, and health sciences. We first consider alternative conceptualizations of “time” beyond developmental time, historical time, and real-time. For example, time could reflect the age of onset for a particular behavior, such as the age at which one first used cannabis. Further still, time could reflect a continuous dimension that is not time at all, for example, to estimate associations as a nonparametric function of a characteristic such as severity of disease. We conclude with a brief summary of future methodological work to expand the capabilities of TVEM, including software extensions, the accommodation of ordinal or zero-inflated Poisson outcomes, and the integration of TVEM and mixture modeling.