Group social skills interventions (GSSIs) are among the most commonly used treatments for improving social competence in youth with ASD, however, results remain variable. The current study examined predictors of treatment response to an empirically-supported GSSI for youth with ASD delivered in the community (N[subscript total]=75). Participants completed a computer-based emotion recognition task and their parents completed measures of broad psychopathology, ASD symptomatology, and social skills. We utilized generalized estimating equations in an ANCOVA-of-change framework to account for nesting. Results indicate differential improvements in emotion recognition by sex as well as ADHD-specific improvements in adaptive functioning. Youth with both co-occurring anxiety and ADHD experienced iatrogenic effects, suggesting that SDARI may be most effective for youth with ASD without multiple co-occurring issues. Findings provide important directions for addressing variability in treatment outcomes for youth with ASD.