Latent Transition Analysis for Modeling Change Over Time: A Demonstration of SAS PROC LTA

Many questions that arise in the fields of public health research are addressed using longitudinal studies. Often, the population in such studies is comprised of a mixture of several sub-populations characterized by different growth processes. Several different statistical models can be used to identify subgroups with different patterns of change over time, depending on the nature of the outcome being modeled. Latent growth mixture models and group-based trajectory models are ideal when change over time in a continuous variable can be characterized as a function of time within each subgroup. In contrast, latent transition analysis (LTA), a longitudinal extension of latent class analysis, can be used to model development in discrete latent variables, for example, stage processes, over two or more times. The author of this study and her colleagues have developed a SAS procedure, PROC LTA, that can be used to model transitions in latent class membership over two or more times. PROC LTA enables researchers to conduct multiple-groups LTA in order to examine measurement invariance across groups and group differences in the prevalence of class membership and transitions over time. Covariates can be included in order to identify important predictors of latent class membership and transitions over time. This study demonstrates how to use PROC LTA to model development in adolescent and young adult dating and sexual risk behavior. Gender differences are examined, and substance use behaviors are included as predictors of initial status in dating and sexual risk behavior and transitions over time.