Latent Transition Models

Flows between states for discrete periods of time can be described with a Markov chain, a process without memory. More realistic for social science data is a heterogeneous mixture of Markov chains, of movers and stayers. Alternatively, or in conjunction with heterogeneity in the tendency to change, one can take measurement unreliability into account, using latent or indirectly measured Markov models. The basic model variants are explained with data examples. References to overview papers are given. Keywords: latent class model; latent Markov model; Markov chains; mover–stayer model