A Bottom-Up Approach To Examining Group-Level Communication Patterns: A Multilevel Latent Profile Analysis Of Functional Group Interaction

Functional perspectives on small discussion groups often focus on either the individual or group level of analysis, but rarely both. Following Borgatta and Bales (1953), I argue that group-level constructs are built from individual actions, and that the distribution of different interaction profiles leads to group-level assessments that are sometimes different from, but based on, assessments made at the individual level. The current study applies multilevel latent profile analysis (MLPA) to a previously published data set to identify interaction profiles at the individual level and determine whether those profiles (a) are interdependent and (b) allow for the classification of groups. MLPA identified three interaction profiles at the individual level and, based on those profiles, two classes of groups emerged. Discussion addresses the impact of the findings on conceptualizing and studying small-group interactions.