Quantitative Analysis of Group Data: Multilevel Latent Variable Models

Multilevel approaches are generally well suited to group communication because what people say and do in groups is a function of intra- and trans-individual mechanisms. This chapter first provides a brief overview of group research as a multilevel problem and then describes more modern approaches to modeling nested data using latent variable models, including multilevel structural equation modeling and latent class analysis. The chapter concludes by addressing conceptual opportunities provided by multilevel latent modeling approaches to group communication.