A multivariate latent growth curve analysis of adolescent substance use

Abstract

This article demonstrates a multivariate latent growth curve methodology (LGM) for analyzing longitudinal adolescent substance use data. Hypotheses concerning the form of growth in substance use, individual differences in the common trajectory over time, and covariates influencing growth were tested. Significant linear increases existed for alcohol, cigarette, and marijuana use. Second-order multivariate extensions of LGM indicated that associations among the individual differences parameters of the various substances could be adequately modeled by a higher order substance use construct. Family status, parent-child conflict, peer encouragement for substance use, parent substance use, and age significantly influenced initial levels of use. Peer encouragement, change in peer encouragement, change in parent-child conflict, age, and gender significantly influenced development of use. These findings support the influence of families and peers on the development of adolescent substance use and demonstrate the utility of multivariate extensions of LGM in the analysis of longitudinal data. © 1996, Lawrence Erlbaum Associates, Inc.