ABSTRACT
This chapter builds on Chapter 7, which demonstrated how a growth mixture model (GMM) can identify unobserved heterogeneity in trajectories with normal continuous variables. The same analytical approach is utilized in this chapter to illustrate how to estimate a GMM with non-normal variables, specifically categorical and count variables (i.e., categorical GMM and count GMMs). This chapter demonstrates how to estimate categorical and count GMMs by incorporating a latent class variable into the categorical and count latent growth curve models (LGCMs) introduced in Chapter 10. The estimation of a categorical GMM (including binary and ordinal) variables is introduced, followed by a count GMM. Further, we briefly describe how these non-normal variables can be used to estimate second-order GMMs.
