The increasing use of diary methods calls for the development of appropriate statistical methods.For the resulting panel data, latent Markov models can be used to model both individual differencesand temporal dynamics. The computational burden associated with these models can be overcome byexploiting the conditional independence relations implied by the model. This is done by associating aprobabilistic model with a directed acyclic graph, and applying transformations to the graph. The structureof the transformed graph provides a factorization of the joint probability function of the manifest and latentvariables, which is the basis of a modified and more efficient E-step of the EMalgorithm. The usefulness ofthe approach is illustrated by estimating a latent Markov model involving a large number of measurementoccasions and, subsequently, a hierarchical extension of the latent Markov model that allows for transitionsat different levels. Furthermore, logistic regression techniques are used to incorporate restrictions on theconditional probabilities and to account for the effect of covariates. Throughout, models are illustratedwith an experience sampling methodology study on the course of emotions among anorectic patients.
展开▼
