Modeling Interaction Effects in Latent Growth Curve Models

This article presents an extension of Joreskog and Yang's (1996) method of estimating interaction effects among latent variables to latent growth curve models. Models involving a product of 2 latent factors of either static or dynamic variables (parental monitoring and rule-setting) predicting an outcome growth variable (adolescents' initial status and rate of change in antisocial behavior) are used as substantive illustrations. The results are discussed in terms of practical and statistical problems associated with interaction analyses in latent curve models, and in structural equation models with latent variables in general.