Linear mixed model trees for ordinal longitudinal data

Ordinal longitudinal data have been receiving increased attention in social sciences, for example to track individuals on daily activities or subjective experiences. In my project, I focus on the statistical analysis of ordinal longitudinal data using ordinal linear mixed models (e.g. Tutz and Hennevogl, 1996), such as the cumulative logit mixed model (e.g. Agresti, 2010). More specifically, I aim at developing a tree-based complementary tool for this model type specifically suited for exploratory analyzes on interaction effects. For example, it is well suited for discovering demographic subgroups that evolve differently across time. Technically, ordinal linear mixed models are combined with model-based recursive partitioning (Zeileis, 2008).