Back to Top

Rise and Shine: The Effect of School Start Times on Academic Performance from Childhood through Puberty

Research suggests that sleep affects pupil learning, as it spurs memory formation and consolidation as well as improves attention and alertness. While pupils’ sleeping patterns are affected by many factors that have little to do with education policy, they are likely to be directly related to one factor entirely determined by such policy: school start time. Since school start times are likely to affect pupils’ sleeping patterns, it is potentially an important tool with which to affect their achievement. Indeed, randomised research suggests that later school start times increase the amount of sleep substantially. For this reason, some have suggested that adolescent pupils should not start school earlier than 11am.

In their paper for the Journal of Human Resources (Autumn 2018), Jennifer Heissel and Samuel Norris analyse the effects of school start times in Florida on pupil achievement, separating the impact among pre-pubescent children and adolescents, which may very well differ due to their different sleeping schedules. A key problem is that school start time is unlikely to be random to pupil achievement but may very well correlate with a range of observable and unobservable characteristics, which in turn affect achievement. To get around this problem, the authors use a clever and novel strategy that takes advantage of how light affects sleeping patterns biologically: less sunlight at night and more sunlight in the morning lead to earlier bedtimes, which in turn leads to more sleep. In this sense, it is school start time relative to sunrise, rather than clock start time per se, that should affect achievement. The authors exploit the fact that there is a sharp discontinuity in sunrise at American time-zone borders, which schools’ clock start times do not adjust for entirely – thereby creating variation in school start time relative to sunrise. By analysing children in Florida who live close to and make relatively short moves across the central-eastern time zone border at some point between the ages of eight and fifteen, while adjusting for fixed pupil differences and school characteristics, they are able to obtain variation in school start time relative to sunrise that is uncorrelated with other factors that affect achievement.

The authors find that the impact of school start times is partly dependent on pupils’ developmental level. While school start time has essentially no impact on mathematics achievement among prepubescent pupils, it has a rather large effect among adolescents, which is detectable from the age of 11 among girls and 13 among boys – which happen to be pretty much exactly the median age of key pubertal transitions for each gender. The results suggests that one hour extra hour of sunlight before school improves math achievement by the equivalent of about 8 PISA points among adolescents. However, in reading, the effect is equivalent to about 6 PISA points for both adolescents and younger pupils. Meanwhile, the authors only find evidence that later school start times reduce absences among younger children, indicating that the mechanism among adolescents has less to do with learning time and more to do with improved alertness and ability to learn.

Certainly, one should note base public policy on just one paper, but other relatively strong research from different contexts – including North Carolina and South Korea at the school level and the US Air Force Academy at the university level – supports the paper’s conclusions. There is also evidence that later school start times have other benefits. For example, research suggests that starting school later decreases late-night teen car accidents significantly, indicating that there are other welfare gains to consider.

Importantly, altering school start times appears to be a rather cheap policy in contrast to many other education interventions that generate similar effects. Since the impact differs for pupils of different ages, it is also possible to target school start times for different year groups in line with the physiological research. Doing so would be a relatively cheap way to improve pupil performance with no pedagogical interventions whatsoever. Overall, therefore, the policy conclusion appears to be a no-brainer: let the kids sleep so their performance can rise.

Gabriel Heller-Sahlgren

This comment piece is also the Editor's Pick in the 2nd issue of the CfEE Monthly Research Digest 2018-19. The piece reviews a paper by Jennifer A. Heissel and Samuel Norris, Rise and Shine: The Effect of School Start Times on Academic Performance from Childhood through Puberty’, published as a manuscript version by SSRN (a free copy of which may be downloaded here).

You can download free copies of back issues of the CfEE Monthly Research Digest here.

Blog Category: 
About the author