Growth Curve Models with Categorical Outcomes

Motivated by the limited available literature on the treatment of longitudinal binary and ordinal outcomes in a growth modeling framework, the goal of this entry is to provide an accessible and practical introduction of this topic for a criminological audience. The parameterization of categorical latent growth models is explained by integrating aspects of the more familiar conventional latent growth models and generalized linear models. Emphasis is placed on the process of model building, evaluation, and interpretation. The entry contains an elaboration of how to include predictors of developmental change in the model for covariate-related hypothesis tests along with remarks regarding of the importance of auxiliary information for assessing model validity and utility. Finally, several model extensions including nonlinear change, generalized growth mixture modeling, and longitudinal latent class analysis are discussed.