Reading focus
Syllabus: Why Do People Harm Others—and What Can Reduce It?
Use the linked corrections when evaluating the administrative conflict estimate. Separate peer influence, perceived norms, recorded incidents, and experienced harm.
Why you’re reading it. Learn why a promising intervention can provide useful evidence while its headline outcome remains uncertain.
A question to consider. A school reports fewer disciplinary incidents after a campaign. What additional measures would distinguish less harm from changes in reporting or enforcement?
Why this reading
Paluck and colleagues connect a theory of social influence to an intervention in school networks. Randomizing schools and invitations to students creates distinct comparisons for changes across schools and influence among peers.
Before you read
A social referent is a student who receives many nominations in the network survey, not necessarily a universally liked student. Intention-to-treat analysis compares assigned invitations even when some invitees do not participate. Administrative reports measure recorded disciplinary events.
Read for
- Separate school assignment, student invitation, attendance, and exposure to participating peers.
- Compare school-level disciplinary reports with students' perceived norms and peer-level outcomes.
- Use the linked corrections when interpreting the headline estimate and its uncertainty; distinguish the size of an estimate from the strength of evidence.
Consider afterwards
- A school wants to adopt the intervention. What additional outcomes and follow-up would show whether fewer disciplinary reports represent a sustained reduction in harm?
Editorial context
The 2018 correction changes the reported reduction from 30% to 25% and describes the overall effect on administrative reports as statistically insignificant at the .05 level, in tension with wording in the original article. The point estimate therefore should not be presented as a firmly established reduction. A 2019 correction adds funding information. The study covers 56 New Jersey schools, with administrative outcomes available for 49. Norm reports, disciplinary records, and experienced harm are different outcomes; no single result establishes every proposed mechanism or transfer to other settings.