A discussion of statistical methods for longitudinal and matched data analysis

An application oriented approach will be taken for this presentation. A brief overview of the nature of intracluster correlation in longitudinal data and matched data will first be given, and the use GEE or mixed-effect models to evaluate the magnitude of the intracluster correlation and to adjust for the effects of the intracluster correlation in the analysis, is then demonstrated.  Some suggestions are provided for data analysts in analyzing matched data.