Estimation and application of the latent group model

Group scholars often describe or conceptualize groups as having characteristics similar to those of individuals. For example, groups can be thought of as motivated or satisfied in degree. Describing groups in such terms assumes that members possess similar levels of the characteristic in question and that latent group-level processes influence convergence on that characteristic. The latent group model (LGM) offers conceptual and statistical means for assessing the structure of latent, group-level factors based on the convergence of variables measured at the individual level. This paper outlines the conceptual issues related to the LGM and offers a detailed example using interaction data from 4-person groups for estimating the model in SAS-SPSS, Mplus, and R syntax are provided in the appendixes. Applications and extensions of the LGM are discussed.