Discipline Reform, School Culture and Student Achievement

New York City public middle schools, all students, three years post-reform

New York City, US; the 2012 reform eliminating suspensions for non-violent, disorderly behaviour.

Difference-in-differences comparing schools more and less exposed to a 2012 New York City policy change that removed suspension as a permissible response to non-violent disorderly conduct. Outcomes are standardized state test scores and the district's school-culture survey. The design answers the objection that discipline reform must trade off the suspended student against the classroom: it measures both.

Math scores in more-affected schools rose 0.05 SD over three years relative to other schools; reading rose 0.03 SD. Only a small portion of the gain is explained by the direct effect on students who would otherwise have been suspended - most of it tracks improvements in measured school culture (student-teacher relationship quality, perceptions of safety). No evidence of trade-offs between students - students benefited even when they were unlikely to be suspended themselves.

Low. Policy-induced variation across schools, with school and year fixed effects; no family selection into treatment.

The cleanest available test of the classroom-order objection to discipline reform, and it goes against the objection. Unreplicated: one city, one reform. Published in The Economic Journal after circulating as an SSRN working paper.