Applied Longitudinal Analysis
*Garrett Fitzmaurice, Harvard School of Public Health
Keywords: Longitudinal, marginal models, GLMMs
In recent years, there have been remarkable developments in methods for longitudinal data analysis. In particular, generalized linear models have been extended to yield two broad classes of models: "marginal models" and "mixed effects models". Both classes of models account for the within-subject correlation among the repeated measures, though they differ in approach. In general, these two classes of models have different targets of inference and therefore address somewhat different questions regarding longitudinal change in the response. In this workshop we will review marginal and mixed effects models for longitudinal data and highlight the main distinctions between them. We will also discuss the types of scientific questions addressed by each of the two classes of models. The main emphasis will be on the practical rather than the theoretical aspects.