The Impact of Restorative Practices on Violence, Psychological Well-being, and Social Cohesion: Evidence from Schools in Bogota
Study Context
Public schools in Bogotá, Colombia, face high rates of peer violence, particularly affecting students displaced by conflict and economic hardship. Nearly 90% of violent incidents reported through the city’s school alert system involve peer-to-peer aggression, with many classrooms serving internally displaced students and Venezuelan refugees. Traditional punitive responses have struggled to break cycles of violence and improve student well-being, positioning restorative practice (RP) as a promising alternative. Restorative practices is an approach that emphasizes strengthening relationships, repairing harm through inclusive dialogue, mutual accountability, and community-building rather than punishment.
In partnership with Bogotá’s Secretary of Education, this randomized controlled trial will test the impact of RP workshops held in classrooms on students’ violence, mental health, social cohesion and learning outcomes.
Study Design
Real-time sampling and randomization will occur until the intervention group includes 150 school-shifts (i.e. full, morning or afternoon shifts) and 150 control school shifts.
The intervention is expected to impact victims and aggressors involved in conflicts and their classmates (9,000 students, assuming 30 students per class). Facilitators will lead four structured sessions on conflict sensitization; harm and shared responsibilities; socioemotional skills for conflict resolution; and a reparation meeting where aggressors commit to repairing harm and preventing recurrence.
Three months after the intervention, surveys will be administered to students and teachers.. Researchers will also use 2023 administrative data—including the city’s school alert system, enrollment records, and school climate surveys—to examine the role of RP in transforming school culture and strengthening support for marginalized groups.
Results and Policy Lessons
Results forthcoming.