|
Abstract:
|
Latent transition analysis (LTA) is designed to estimate the probability of individuals' membership in hypothesized latent classes and their transition rate among these classes. LTA derives from the family of latent class analysis in which manifest items are treated as fallible indicators of unseen states. Routines for maximum-likelihood (ML) estimation are currently available in statistical software. In many examples, however, the likelihood function exhibits unusual features, causing ML estimates to behave erratically. In this talk, we explore a variety of theoretical and practical issues surrounding the use of the LTA model, including Bayesian alternatives to ML estimation. We illustrate difficulties in ML and Bayesian methods with an example from substance use behaviors in adolescent females.
|
