Can Restorative Practices Improve School Climate and Curb Suspensions? An Evaluation of the Impact of Restorative Practices in a Mid-Sized Urban School District

44 schools in the final sample, evenly split between treatment and control, randomised within 23 matched pairs (2 assigned schools were withdrawn for capacity reasons before the trial)

Pittsburgh Public Schools, US; grades K-12, two years of implementation, 2015-16 to 2016-17.

One of the first randomised trials of restorative practices — the IIRP SaferSanerSchools Whole-School Change programme, run locally as Pursuing Equitable and Restorative Communities (PERC). Randomisation within matched pairs; outcomes are administrative (suspension days, suspension incidence, arrests, mobility, attendance) plus state standardized tests (PSSA) and district climate surveys. Evaluated by RAND under a Department of Justice grant, independent of IIRP. High-school estimates are invalid because of baseline non-equivalence and are excluded by the authors.

Suspensions fell: days lost to suspension declined 36% in PERC schools versus 18% district-wide, and the share of students suspended fell to 13% in PERC schools versus 15% in control schools. Disparities by race and income narrowed. But the effect was driven entirely by ELEMENTARY schools; there was no suspension reduction in middle grades, none for male students, none for students with IEPs, none for violent or weapons incidents, and no reduction in arrests. Teachers rated school climate significantly better; STUDENT-rated classroom climate (Tripod) was LOWER in PERC schools overall. And achievement went the wrong way: the primary combined math-and-reading state assessment estimate was -0.065 (marginal), math alone -0.068 (significant), reading -0.055 (ns), with significant negative effects concentrated in grades 6-8 and in predominantly African American schools. The authors note the suspension benefit and the achievement harm land in DIFFERENT grade bands, so lost instruction from suspension cannot be the mechanism.

Low. School-level randomisation within matched pairs.

Mixed. A second independent cluster-randomised trial of restorative practices (Huang, Gregory & Ward-Seidel 2023, 18 schools) found no main effect on suspension at all, and a two-year cluster-randomised trial in Maine (Acosta et al. 2019, 13 schools) found no effect on student-perceived school climate. The reproducible part of the restorative-practices literature is the elementary suspension reduction; nothing else has held up twice.