Skip to content
The research library
Shift work and healthEmerging evidence

How Good Is the Evidence That Daylight at Work Improves Sleep?

The 46 minute figure quoted everywhere comes from actigraphy on 21 people, and knowing that changes what you should spend on it.

Reviewed against primary sources on September 8, 2026 by the Soon operations research team. How we vet the evidence

The evidence in one line

In a case-control pilot study of 49 office workers, those with windows received 173% more white light during work hours and slept 46 minutes longer per night, with worse sleep quality, vitality and physical-role scores among the windowless group (Boubekri et al., 2014). The objective sleep measurement came from actigraphy in a subset of 21 people. The direction is consistent with wider seasonal-light evidence, but the sample makes the 46 minute figure an estimate to test rather than a number to plan against.

What the study found, and on how many people

Office workers in windowless workplaces were compared with workers in comparable roles in daylit ones. Those with windows received 173% more white light during work hours and slept 46 minutes longer per night; the windowless group reported poorer overall sleep quality and more sleep disturbance, and scored worse on quality-of-life measures covering vitality and physical role limitations (Boubekri et al., 2014).

The design is a case-control pilot. Of the 49 participants, 27 worked without windows and 22 with them, and the objective sleep measurement came from wrist actigraphy in a subset of 21. A 46 minute difference derived from 21 people is a real observation and a fragile estimate, and the two descriptions are not in tension. This is the kind of study that justifies a larger one, which is what its own authors call it.

Why the direction is more credible than the magnitude

The mechanism has independent support. Among 32 indoor day workers at 67.86 degrees north, winter sleep onset came 39 minutes later than in summer and sleepiness was higher, and greater morning light exposure was associated with earlier mid-sleep while greater evening exposure was associated with later timing (Lowden et al., 2018). A year-long study of 30 office workers at 56 degrees north found seasonal variation in positive affect, sleep-activity behaviour, time spent outdoors in daylight and appraisal of workplace lighting (Adamsson et al., 2018).

So three separate groups, none large, point the same way: workplace light exposure tracks with sleep timing and self-reported wellbeing. What none of them establishes is a dose. No study here tells you how many minutes of sleep a given desk position buys, and none measured work output at all, so a business case resting on a specific sleep gain is resting on the weakest part of the evidence.

How to use evidence at this strength

Small consistent studies are good grounds for a low-cost, low-regret change and poor grounds for an expensive one. Rearranging desks, moving a break area toward a window, or protecting a daylight break during winter cost little and align with the direction of every study cited here. Justifying a building retrofit on a 46 minute sleep gain does not, because that number rests on 21 people and has not been replicated at scale.

The stronger move is to treat your own workplace as the next study. Light exposure is measurable, sleep is measurable with the same wrist devices these studies used, and absence and lateness are already in your systems. A local before-and-after with a comparison group produces evidence about your building, which is the only building the decision concerns.

What this means for your schedule

  • Treat the 46 minute figure as a pilot-scale estimate from 21 actigraphy participants, not a planning number.
  • Make the cheap changes the evidence direction supports, such as desk placement, break-area light and protected daylight breaks in winter, without promising a specific sleep gain.
  • Do not fund a lighting or building investment on this evidence alone; require either a local trial or larger intervention studies first.
  • If you do run a local trial, measure light exposure and sleep objectively and keep a comparison group, because self-report alone will not settle it.

The business case

The direction of this evidence is consistent across three independent groups, which justifies low-cost changes now and a measurement plan before anything expensive.

Vendors and design guides frequently quote the 46 minute figure without its sample size; knowing the study had 21 people under actigraphy is what makes a procurement conversation honest.

Frequently asked questions

How many people were in the windows and sleep study?
Forty-nine office workers took part, 27 in windowless workplaces and 22 in daylit ones, and objective sleep was measured by actigraphy in a subset of 21 (Boubekri et al., 2014). It is described by its authors as a case-control pilot study, which is the appropriate weight to give it.
Does this prove windows cause better sleep?
No. It is a case-control comparison, not a randomised trial, so the groups may differ in ways beyond window access, including job type, seniority, commute and time spent outdoors. The consistent direction across separate studies strengthens the case for the mechanism without establishing that the window itself caused the difference.
Is there stronger evidence for the underlying mechanism?
The link between light timing and sleep timing has support beyond this study. Among Arctic indoor day workers, greater morning light exposure was associated with earlier mid-sleep and greater evening exposure with later timing (Lowden et al., 2018). That supports light as the operative cue, though it is also a small field study rather than an intervention trial.
What about productivity rather than sleep?
None of these studies measured work output. They measured light exposure, sleep timing and duration, sleepiness, mood and quality-of-life scores. Any step from those outcomes to productivity is an inference the studies do not make, and figures presenting daylight as a percentage output gain are not coming from this literature.
What would better evidence look like?
A randomised or stepped-wedge intervention in real workplaces, with a few hundred participants, objective light and sleep measurement, and pre-specified work outcomes such as absence, lateness or error rates. Until something like that exists, the honest position is a plausible direction with an unestablished effect size.

Sources

Every figure on this page is drawn from a cited primary source and checked against the original publication.

  1. Boubekri, M., Cheung, I. N., Reid, K. J., Wang, C.-H., & Zee, P. C. (2014). Impact of Windows and Daylight Exposure on Overall Health and Sleep Quality of Office Workers: A Case-Control Pilot Study. Journal of Clinical Sleep Medicine, 10(6), 603โ€“611. https://doi.org/10.5664/jcsm.3780

    Design: Case-control pilot study of 49 office workers, with actigraphy in a subset of 21

  2. Lowden, A., Lemos, N. A. M., Gonรงalves, B. S. B., ร–ztรผrk, G., Louzada, F., & Pedrazzoli, M. (2018). Delayed Sleep in Winter Related to Natural Daylight Exposure among Arctic Day Workers. Clocks & Sleep, 1(1), 105โ€“116. https://doi.org/10.3390/clockssleep1010010

    Design: Within-person winter and summer field study (32 office workers at 67.86ยฐN)

  3. Adamsson, M., Laike, T., & Morita, T. (2018). Seasonal Variation in Bright Daylight Exposure, Mood and Behavior among a Group of Office Workers in Sweden. Journal of Circadian Rhythms, 162. https://doi.org/10.5334/jcr.153

    Design: One-year longitudinal field study (30 office workers at 56ยฐN)

Cite these sources: BibTeX RIS

Why this page is graded emerging evidence

A single cohort, field, or simulator study. Worth knowing, not yet worth building policy on.

Who reviewed this

Every article in this library is checked against its primary sources by the Soon operations research team: each figure is traced back to the study it came from, and the wording is checked against the study design before publication. What that review covers

None of the studies cited here evaluated Soon.They examine scheduling practices, shift patterns, and working hours as studied by independent researchers, so their findings describe what those practices are associated with, not what any particular software produces.

This article summarizes published research for scheduling and operations decisions. It is not medical advice. Individual health questions belong with a qualified clinician.

Your next schedule could take 2 minutes.

Import your team, set your rules, hit auto-fill. Most teams are live the same day.

Try Soon free

30 days free ยท No credit card required

Already have an account?Sign in