A Randomized Controlled Trial of Restorative Justice in High-Suspending Brooklyn Schools

• In Event: 21.062 - MTCC Roundtable Session Six
  In Roundtable Session: 21.062-6 - Restorative Justice in Schools: Rationale, Implementation, Outcomes, and Caveats

Abstract

Recognizing the potentially deleterious consequences of criminalizing school discipline, schools are increasingly turning to alternative methods for holding students accountable for misbehavior. Restorative justice (RJ) practices—which seek to hold students who cause harm accountable without removing them from their learning environment—ostensibly represent an antidote to traditional discipline. However, RJ practices have been the subject of limited high quality scientific inquiry. This ongoing research project, in partnership with the New York City Department of Education, aims to fill this gap through the implementation and evaluation of a restorative program in a school district that struggles with the highest suspension rates in New York City, as well as elevated neighborhood crime, including high rates of gun violence and gang activity, where the need for effective school disciplinary practices is greatest. The study involves high school students from 10 schools in District 18 of New York City, as well as students in similar schools from three other nearby districts. Utilizing a randomized controlled trial, 10 schools have been randomly assigned to either receive the program or to the control group. Because of the DOE’s ongoing work in the district, the evaluation will use a three-group design that will compare outcomes among: (1) randomly assigned District 18 schools that will implement the peace circle program; (2) control group schools in District 18 that will not implement the peace circle program but will still receive general restorative practices training; and (3) matched comparison schools from nearby school districts in Brooklyn. The study will involve an process evaluation, including fidelity analysis, and an impact evaluation. Official data will be obtained from the NYC Department of Education. After data reduction, weighting, and descriptive analyses, hierarchical linear modeling will be utilized for two distinct statistical modeling approaches. The first will solely involve school-level outcomes—and all predictors will be at the school-level only. The second approach will involve individual-level data. This presentation aims to present preliminary findings from Year 1 of the study and shared lessons learned and research-practice successes.

Author

• Andrew Martinez, Center for Court Innovation