Latent Growth Mixture Models
DOI link for Latent Growth Mixture Models
Latent Growth Mixture Models book
The classes of a categorical latent variable (c) can be proposed to represent latent trajectory classes, which classify individuals into sets of exclusive categories (Clogg, 1995). The multiple indicators of the latent classes correspond to repeated measures over time (Muthén, 2002). Within-class variation is permitted for the latent trajectory classes (Muthén & Muthén, 1998-2012). This variation is also represented by random effects that influence the outcomes at all time points. Covariates can be added to the model, both to describe the formation of the latent classes if desired and how they may be differentially measured by the repeated measures (Asparouhov & Muthén, 2013). As in previous examples, the prediction of latent class membership is by the multinomial logistic regression of c on x.