School-based interventions to reduce suspension and arrest: A meta-analysis

Objectives: This meta-analysis summarizes recent experimental evidence on school-based interventions designed to reduce student suspensions and arrests. Methods: Eligible studies included randomized controlled trials from 2008 to 2019 with at least 100 students and an official measure of either suspension or arrest. Fourteen studies were found, resulting in 12 suspension effects and 6 arrest effects. Results: Overall, interventions showed small but not statistically significant reductions in arrests and suspensions. The weighted mean of the arrest effects was Cohen's d = -0.013 (p = .093) and the weighted mean of the suspension effects was d = -0.033 (p = .215). Suspension effects differed by intervention characteristics: well implemented interventions demonstrated significant reductions in student suspensions (d = -0.11, p < .001) but interventions with explicit implementation problems did not (d = 0.047, p = .084). Interventions in high schools reduced suspensions significantly (d = -0.137, p = .003) whereas elementary school interventions did not (d = -0.021, p = .251). Conclusions: Well-implemented school-based interventions may reduce suspensions by building positive environments and providing supports for students. More research should be done to identify ways in which these and other programs can best be implemented to reduce suspensions and arrests.