Modeling joint and marginal distributions in the analysis of categorical panel data

This article presents a unifying approach to the analysis of repeated univa riate categorical (ordered) responses based on the application of the gener alized log-linear modeling framework proposed by Lang and Agresti. It is sh own that three important research questions in longitudinal studies can be addressed simultaneously. These questions are the following: What is the ov erall dependence structure of the repeated responses? What is the structure of the change between consecutive time points? and What is the structure o f the change in the marginal distributions? Each of these questions involve s specifying log-linear models for different marginal distributions of the multiway cross classification of the responses. The proposed approach is il lustrated by means of two real data examples.