Seattle Later School Start Times — Natural Experiment
University of Washington / Salk Institute · Seattle, WA, USA · 2016
Summary
The Seattle later-start experiment demonstrates the simplest possible public health intervention: do nothing more than change when an existing institution opens its doors. Adolescent sleep biology is well established — puberty shifts the circadian clock so that teenagers naturally fall asleep later and wake later. Starting school at 7:50am asks teenagers to be awake and cognitively functional at hours when their biology fights against it. The Seattle data showed that 55 additional minutes of school start time translated to 34 additional minutes of actual sleep per night, sustained across the academic year. The resulting improvements in attendance, grades, and — most strikingly — car accident rates show that the sleep gain was real and consequential. The policy is almost free to implement (it requires only schedule changes, not additional staff or curriculum), making it one of the highest benefit-to-cost public health interventions available to school districts.
Research question
"Does delaying high school start times by ~55 minutes increase sleep duration and improve academic performance, attendance, and health outcomes in adolescents?"
Methodology
Intervention
In fall 2016, Seattle Public Schools delayed start times for two high schools from 7:50am to 8:45am. Two other high schools maintained the earlier schedule. Researchers measured student sleep via wrist-worn actigraphy devices before and after the policy change, comparing schools that changed start times to those that did not. The study, published in Science Advances, is the strongest pre-post comparison available for this question in the US.
Assignment
Natural experiment with difference-in-differences design; two treatment schools (delayed start) vs. two comparison schools (unchanged start); student sleep tracked with actigraph wristbands pre- and post-change; academic and attendance records from Seattle Public Schools
Sample size
~180 students with complete actigraphy data; school-level academic records for 1,000+ students per school
Primary outcome
Nightly sleep duration; school attendance rates; grade point average; teen car accident rates (state-level analysis)
Effect estimate
Sleep: students slept 34 more minutes per night after the delay (gain sustained through the school year, not just the first weeks). Attendance: significantly improved in treatment schools vs. comparison. GPA: median grade point average improved in treatment schools. Teen car accidents: state-level analysis found 6 fewer accidents per year per 1,000 licensed teen drivers in districts with later start times.
Decision
California enacted legislation in 2019 requiring all middle and high schools to start no earlier than 8:00am and 8:30am respectively, effective 2022 — the first state-level mandate in the US; American Academy of Pediatrics, American Medical Association, and American Academy of Sleep Medicine all recommend later start times; districts in multiple states have voluntarily delayed start times; transportation costs are the primary barrier to adoption; Crain et al. (2020) found broader health effects including lower rates of depression and obesity in districts with later starts
Result
Positive
Sleep: students slept 34 more minutes per night after the delay (gain sustained through the school year, not just the first weeks). Attendance: significantly improved in treatment schools vs. comparison. GPA: median grade point average improved in treatment schools. Teen car accidents: state-level analysis found 6 fewer accidents per year per 1,000 licensed teen drivers in districts with later start times.
Evidence strength
Moderate
Quasi-experimental design; causal interpretation requires care.
Replication status
Partially replicated
Institution
University of Washington / Salk Institute
Location
Seattle, WA, USA
Year
2016
Policy area
Education
Mechanism
Default