Although some literature reviews cited mixed results for group-based social competence interventions in schools, existing descriptions of intervention efficacy often lack attention on important factors that may moderate differential response. Some studies suggest that individual characteristics such as diagnosis (i.e., specific deficit clusters) or behavioral profiles may be important social outcome moderators. Given the interactive nature of group-based interventions, understanding how individual characteristics combine to influence outcomes for the group as a whole is an important next step. By using a multisite cluster randomized control trial (n = 274), the current study explores the impact that varying student characteristics have on outcomes of the Social Competence Intervention for Adolescents (SCI-A; Stichter, Herzog, Owens, & Malugen, 2016; Stichter et al., 2010) compared to outcomes of business-as-usual (BAU) practices. Researchers hypothesized that (a) SCI-A would be more effective than BAU in improving ratings of students' social functioning when students within groups were more similar to each other on relevant diagnostic and behavioral indicators, and (b) that within SCI-A only, more versus less homogenous groups would demonstrate greater social outcome gains. Results offer some support for these hypotheses: SCI-A was more effective in improving social communication and motivation when groups were diagnostically similar (ds > 0.55). Considering changes in social awareness and communication, SCI-A was more effective when groups were heterogeneous on socially competent behavior (ds > 1.32) and less effective when groups were heterogeneous on antisocial behavior (ds > 1.00). The authors discuss the implications and importance of research exploring factors such as group composition that may moderate intervention response in applied settings.