Abstract
Outcomes from a large cluster-randomized study of Second Step©, a commonly adopted universal SEL program, have previously demonstrated small, teacher-reported effects on social–emotional competencies for subgroups of students. Given the size of investment in large RCT studies, and in SEL programming, replication attempts are warranted, ideally with diverse analytical strategies that can build a more convincing body of knowledge. The current manuscript builds upon previous study findings by utilizing a growth mixture modeling approach on a more limited set of outcomes. Intervention differences were significant for identified classes of hyperactivity, conduct problems, and emotional symptoms. Patterns suggest that Second Step© plays a predominantly mitigating role for those with modest levels of conduct problems and hyperactivity. Additionally, students in Second Step© schools were more likely to experience decreasing levels of emotional symptoms, as well as mitigation of escalation of symptoms. No academic effects were found, nor effects on prosocial skills. Findings are largely consistent with previous but different analytic approaches but extend to conduct problems and help in elucidating how universal SEL programming works. Implications for Second Step© implementation and future studies are briefly discussed.