Yet the same theoretical structure that agrees with these recent observations, makes some surprising predictions. Namely: that the effectiveness of ‘countermeasures’ such as delaying social schedules without concerted changes in an individual’s light consumption behavior is likely to be limited; that endeavoring to improve sleep timing by shifting alarm time will necessarily lead to a further delay of the circadian sleep propensity rhythm; that delaying social schedules is relatively more effective in individuals who are already early types (shorter intrinsic periods) or on schedules that are very early relative to solar time (e.g., US school schedules) so that those who are affected the most by social jetlag will be benefitted the least; that a relatively small reduction in evening light is a similarly effective intervention to changing the alarm from 7:00?h to 8:00?h; that the reduction in the rate of homeostatic rise during wake across adolescence35 renders adolescents particularly sensitive to the effect of evening light, and that in hunter-gatherer societies without access to evening light, large differences in sleep timing across the lifespan will not be observed.

If, in hindsight, these predictions seem obvious, it should be remembered that much of the literature on preferred sleep timing, so-called chronotype, makes no mention of light. Consequently, chronotype is often reported as if it were trait-like rather than a combination of intrinsic and extrinsic factors31. This then leads to the misconception that adolescents are ‘programmed’ to wake up late. In the debate on changing school start times, there is little mention of light consumption behavior or the light environment. Clock time is used as a measure of whether a start time is ‘early’, yet some time zones span many degrees of longitude so what is ‘early’ on the Western edge of a time zone will not be so early on the East. Adolescents may also experience very different day-time new-lights levels depending on their geographical location. Consequently, changes to school start times seem ill-advised without full consideration of the light consumption behaviour. It may be that the different conclusions reached55 on the relative effectiveness of moving school start times can be explained by differences in the light environment.