ABSTRACT

This chapter draws on the work of Kenny and Judd, extending it to the analysis of more complex models with interactions among growth parameters. It presents an example of an latent growth modeling (LGM) involving interaction effects. The method is an extension of conventional nonlinear structural equation modelings to LGMs where the primary focus is on modeling interactions between latent growth parameters. However, in the context of LGMs where latent means are an important component of the growth parameters, forcing means to be zero may complicate parameter estimation. Given the recency of the numerical integration approach, its utility in analyzing interactions in LGMs is unknown. Although there has been an increasing application of LGMs in the social sciences, LGMs with interactions representing different models of change have been conspicuously absent despite their promise for testing complex hypotheses in both cross-sectional and longitudinal studies. The method presented can be a useful tool for detecting whether interactions between interindividual growth parameters improve prediction of behavioral outcomes.