University of Washington - Department of Statistics
Latent class transition models (LCTMs) are used to study the movement of individuals among homogeneous subgroups through time. Traditional LCTMs assume a complete set of observations for each individual. However, many longitudinal surveys have a rolling enrollment design, with late entry and early exit. Thus, methodology is needed to account for all the possible times at which individuals can be observed.
In this talk, I develop a group-based modeling approach that encompasses these observation paths, and extend estimation algorithms for such models to incorporate covariates in multiple model components. I comment on the structural similarity between LCTMs and Hidden Markov Models, and compare methods used to compute standard errors within both modeling frameworks. I then illustrate this methodology using chronic disability data from the National Long-Term Care Survey (1984-2004), and examine the recent evidence for disability decline. In addition, I explore gender differences in both class composition and transition rates among classes, with special emphasis on transition rates into death.