POPULATION-AVERAGED AND CLUSTER-SPECIFIC MODELS FOR CLUSTERED ORDINALRESPONSE DATA

We compare population-averaged and cluster-specific models for cluster ed ordinal data. We consider generalized estimating equations and cons trained equations maximum likelihood estimation of population-averaged cumulative legit regression models, and mixed effects estimation of c luster-specific cumulative logit regression models. A previously repor ted relationship between population-averaged and cluster-specific para meters for the binary logistic link appears to hold for analogous para meters under the cumulative logit link. We address these issues in the context of data from two cross-over clinical trials.