Answering the Bell: High School Start Times and Student Academic Outcomes

Statewide student-level administrative data, all North Carolina public high schools

US public high school students, North Carolina, including traditionally disadvantaged subgroups

Statewide student-level analysis with rich controls, but the variation exploited is cross-sectional differences in start times between schools rather than a policy shock. The authors themselves report "associations", not effects. They also document that urban high schools are likely to start either very early or very late, which is precisely the sorting pattern that makes cross-school comparisons treacherous — start time is correlated with district resources, transport systems, and neighbourhood composition.

Later start times were associated with positive ENGAGEMENT outcomes — reduced suspensions and higher course grades — with larger associations for disadvantaged students. ACHIEVEMENT results were mixed, with both positive and negative associations between start times and test scores. The authors close by calling for further research to evaluate whether later start times are a scalable, cost-effective approach to boosting engagement and achievement, i.e. they do not treat their own estimates as settling it. The split between a consistent behavioural signal and an inconsistent test-score signal is the same split that runs through the whole literature.

High. Start times are not randomly assigned across North Carolina schools; urban schools cluster at the extremes, and family sorting into schools by start time and by neighbourhood is unaddressed. Course grades are also assigned within the same schools whose characteristics drive the start time. Nothing here rules out selection.