THE ANALYSIS OF DESIGNS IN WHICH INDIVIDUALS ARE IN MORE THAN ONE GROUP

The design and analysis of data when individuals are in more than one group are discussed. In the generations design, the group members chan ge across time in a pattern that mimics the birth and death of members in families. We present a regression-based approach to the estimation of conformity and stability parameters. In a rotation design, people interact in multiple groups and multiple sessions and each group invol ves a different set of partners. For two different types of rotation d esigns, we present a variance decomposition approach. For all designs discussed, we illustrate the analysis with results from actual studies .