On Analyzing Dependent Data

Abstract

In this talk I will discuss some issues concerning the analysis of dependent data. Both single series and multiple short series will be addressed with an emphasis on the importance of an appropriate dependence structure. Model selection based on the predictive power will be stressed. Special emphasis will be on the analysis of growth curve data. One of the major themes is that when faced with real growth curve data, we should keep an open mind regarding the selection of a covariance structure corresponding to fixed, mixed, random effects, or other more sophisticated models.