What Works to Improve Social-Emotional Skills in Conflict and Crisis Settings: Evidence from a Meta-Analysis of Education Interventions

This meta-analysis synthesizes findings from 20 rigorous evaluations of seven types of education interventions from multiple conflict and protracted crisis settings to assess their effects on children's social-emotional learning (SEL) outcomes, in the social, emotional, cognitive, values, perspectives and identity domains: Mental Health and psychosocial support (MHPSS), academic support with positive school climate practices, SEL-targeted practices, comprehensive SEL curricula, remote academic support through EdTech, multi-component interventions, and peace education. We estimate a pooled impact of 0.10 standard deviations (SD), reflecting medium, positive and statistically significant effects. Domain-specific analysis shows medium, positive and significant effects in the emotional (0.13 SD), social (0.09 SD) and perspectives (0.07) domains, but limited evidence and null effects in the cognitive, values and identity domains. Across intervention types, academic interventions that integrate SEL components -through positive school climate, SEL-targeted practices or comprehensive SEL curricula- show significant, small to medium effects (0.05 to 0.13 SD), and MHPSS interventions exhibit the largest impacts (0.16 SD), but effects are not statistically significant given high variability. Analysis combining cost with impact data suggests that while education interventions can improve SEL, their cost-effectiveness varies considerably, and many do not achieve strong cost-effectiveness. Compared to high- and middle-income contexts, smaller effect sizes in conflict and protracted crisis settings may reflect challenges in implementation, contextual fit, and measurement. The study underscores the importance of designing interventions to support SEL outcomes in humanitarian settings that are developmentally appropriate, contextually grounded, and scalable from the outset.