Last updated: October 2026
No screens an hour before bed is the most repeated rule in sleep advice, and almost nobody says what to do instead. The short answer: protect the bedtime, not the darkness — and build the hour around getting into bed on time.
Key takeaways
- Among 122,058 adults in the American Cancer Society’s Cancer Prevention Study–3, daily screen use in the hour before bed went with a bedtime 18.8 minutes later on workdays, 7.6 minutes less sleep, and a 33% higher prevalence of self-rated poor sleep quality (Zhong et al., JAMA Network Open, 2025).
- The cost sits mostly in displaced time, not in the light: across 11 experimental studies, bright screen light delayed falling asleep by an average of 2.7 minutes (Reynolds and colleagues, summarising their own review).
- Melatonin suppression appears to need roughly 1,000–2,000 lux. A phone or laptop screen emits about 80–100 lux; daylight is around 100,000 lux.
- A 16-person National Sleep Foundation panel voted on 10 statements about screens and sleep and reached consensus on 5. Every agreed statement naming an age group named children and adolescents, not adults (Hartstein et al., Sleep Health, 2024).
- A warm bath or shower at 40–42.5 °C (104–109 °F), finished 1–2 hours before bed, went with falling asleep about 10 minutes faster in a review that screened 5,322 studies (Haghayegh et al., Sleep Medicine Reviews, 2019).
- Evening types pay more for the same habit: 15.6 minutes later to bed against 9.3 minutes for morning types.
- The rule that survives the evidence is narrower than “no screens”: nothing in that hour may be open-ended.
What does the evidence actually say about screens in the hour before bed?
It says the habit costs you time in bed — a small amount per night that adds up over a week. The 2025 analysis in JAMA Network Open asked 122,058 adults almost exactly the question people argue about online, and compared daily users against people reporting none.
Daily users went to bed 18.82 minutes later on workdays and 19.69 minutes later on non-workdays. Sleep duration fell by 7.64 minutes on workdays and 5.04 on non-workdays. Across seven nights that is roughly 50 minutes of sleep. The prevalence of poor self-rated sleep quality was 33% higher among daily users than among people reporting no screen use, and the pattern was graded: 1–3 days a week of use sat at 11% higher, 4–6 days at 14%.
Two limits belong next to those numbers. The study is cross-sectional, so it cannot separate “the screen delayed bedtime” from “late sleepers have more time to use a screen”. And sleep was self-reported, though the authors validated their duration measure against sleep diaries and actigraphy in the same cohort.
Is it the light, the content, or the clock?
Mostly the clock. The three mechanisms usually offered have been tested separately, and they come out very differently in size.
| Mechanism | What was measured | Effect found |
|---|---|---|
| Bright screen light | 11 experimental studies, time to fall asleep | 2.7 minutes later on average; some studies found better sleep after screen use |
| Arousing content | 7 studies, alerting content (games) vs calmer content (TV) | about 3.5 minutes less sleep |
| Displaced time | Device use continuing past the point sleep was possible | the mechanism where real sleep loss shows up |
Those figures come from the research team’s own summary of their review. The light explanation runs into a dosing problem: the published threshold for a meaningful melatonin effect is around 1,000–2,000 lux, and a device screen puts out roughly 80–100 lux at the eye. Our explainer on how blue light from electronic devices affects sleep covers why the laboratory conditions that do suppress melatonin look nothing like a phone in a dark bedroom.
The consequence is a reordering, not a dismissal. If light costs minutes and displacement costs tens of minutes, a night-mode filter solves the smaller problem and an alarm marking the start of the hour solves the larger one.
Where does the “one hour” rule come from?
Not from a trial comparing one hour against thirty minutes. It comes from expert guidance, and that guidance is more cautious than its popular version.
The National Sleep Foundation convened a 16-person panel in 2023 for exactly this question. It narrowed 2,209 articles to 522 empirical papers and 52 reviews, then to 35 experimental studies capable of supporting a causal claim, and voted twice on 10 candidate statements using a modified Delphi method. It reached consensus on 5 of the 10.
The agreed statements are worth reading closely: screen use in general impairs sleep health among children and adolescents, pre-sleep screen content impairs sleep health for the same groups, and behavioural strategies may reduce the negative effects. The adult picture did not produce the same agreement — so the rule is routinely handed to adults with more confidence than the panel used.
That does not make the hour useless. It makes it a planning device rather than a prohibition, which fits the displacement data better.
Does a Kindle, a TV across the room, or an audiobook count as a screen?
Judge each one by whether it can keep you up, not by whether it emits light. Once light is worth minutes and open-ended time is worth tens of minutes, the question “is this a screen?” becomes less useful than “does this end by itself?”
| What you are doing | Light at the eye | Can it run past your bedtime? | Practical reading |
|---|---|---|---|
| E-ink reader, front light low | Very low | Only if the book is gripping | Closest thing to a paper book |
| Backlit tablet, one chapter | Low in absolute terms | Yes, but a chapter has an end | Usually fine if the stopping point is set first |
| Phone, social feeds | Low | Yes — by design, there is no end | The high-risk case |
| TV across the room, one episode | Low at that distance | Yes, autoplay removes the end | Fine with autoplay off |
| Audiobook or podcast, sleep timer on | None | No, the timer ends it | Low risk; keep the volume low |
| Video games | Low | Yes, and the content is the alerting kind | Both mechanisms at once |
The stopping point matters more than the screen. A paper book you cannot put down displaces sleep as effectively as a feed; the difference is that a feed is built to have no natural ending. This is why “I only read on my Kindle” and “I only watch one episode” tend to hold up — both come with a built-in end.
4 things that decide whether your wind-down hour works
- Whether the hour has a fixed start. A wind-down that begins “when I feel tired” begins after bedtime. The displacement finding above is a finding about start times.
- Whether the activity ends on its own. Chapters, episodes with autoplay off, sleep timers and bath water that goes cold all end. Feeds, recommendation queues and multiplayer games do not.
- Whether you are in bed during it. Spending the hour in bed awake is the one choice that works against you, because it trains the association between bed and being alert. Reserving the bed for sleep is the behavioural instruction known as stimulus control, which the American Academy of Sleep Medicine’s 2021 guideline carries as a recommended single-component therapy and as one of the components inside cognitive behavioural therapy for insomnia.
- Your chronotype. The same habit cost evening types 15.62 minutes of bedtime against 9.33 for morning types in the JAMA Network Open analysis, and 8.36 minutes of sleep against 5.64. If you are a late type, the hour is doing more work and is worth defending harder. Our guide to circadian rhythms and how they influence sleep explains why the same clock time is not the same biological time for everyone.
What can you actually do for an hour that has evidence behind it?
Four options have something more specific than “it relaxes you” behind them.
A warm bath or shower, finished 1–2 hours before bed. Biomedical engineers at the University of Texas at Austin screened 5,322 studies and pooled the usable ones. Water at about 40–42.5 °C (104–109 °F), timed 1–2 hours before bedtime, was associated with falling asleep roughly 10 minutes faster. The proposed mechanism is heat dissipation: warming the skin opens peripheral blood vessels, which helps core temperature fall on the schedule the circadian rhythm expects. A bath immediately before bed is a different intervention from the one that was studied.
Dimming the room rather than policing the device. If a meaningful melatonin effect needs something near 1,000–2,000 lux, overhead lighting is a more plausible target than a phone held at reading distance. Lamps instead of ceiling lights cost nothing to test.
Staying out of bed until you intend to sleep. The least appealing option, and the one with the most clinical support: the AASM guideline above gives cognitive behavioural therapy for insomnia its one strong recommendation, and stimulus control is part of it. The bed is used for sleep; the hour happens in a chair. If you are already lying awake for long stretches, our overview of why you can feel tired after a full night’s sleep is a useful companion.
Low-stakes reading, with the stopping point chosen first. The evidence here is thinner than for the bath, and the honest version is that reading works because it ends, not because of anything special about paper.
How do you build an hour before bed without screens that you will keep?
Start from the alarm, not from the evening. Pick the wake time you actually have to keep, subtract the sleep you need, and the result is bedtime. Subtract another hour and you have the start of the wind-down — which is a time, not a feeling.
Then do two things with it. Put the open-ended activity before that start time rather than after it, because the point is not abstinence but position in the evening. And set the alarm on the hour’s start, not on the bedtime: by the time a bedtime alarm goes off, the displacement has already happened.
Our practical guide to better sleep covers the room the hour ends in; why you change sleep positions at night and the guide for side sleepers cover what happens once you are down.
When is a bad wind-down hour not the problem?
When the pattern does not match the behaviour. A wind-down hour addresses a bedtime that keeps drifting later. It does not address falling asleep quickly and waking at 3 a.m., waking unrefreshed after eight uninterrupted hours, loud snoring with witnessed pauses in breathing, or months of lying awake despite a stable schedule. Those patterns may point to something a schedule change will not reach, and they are worth raising with a doctor.
The same caution applies in reverse. Eighteen minutes of delayed bedtime is a population average, and averages describe groups. If your own hour is screen-heavy and your sleep is fine, the honest reading of this evidence is that you are not the person the warning was written for.
Methodology: this guide is based on published research and expert-panel statements cited inline, not on product testing.
Frequently asked questions
Do I really need a full hour with no screens before bed?
Probably not a full hour, and probably not literally no screens. The largest adult study compared daily screen use in the hour before bed against none, and found about 19 minutes of delayed bedtime — a timing effect. What the evidence supports is a fixed point in the evening after which nothing open-ended starts, which is a narrower rule than total abstinence.
Does reading on a Kindle count as screen time?
For light purposes, an e-ink reader with a low front light is close to a paper book, and even a backlit tablet sits far below the 1,000–2,000 lux that appears to be needed for a meaningful melatonin effect. For bedtime purposes it counts if it keeps you up. A chapter ends; a browsing session does not.
Is watching TV before bed better than scrolling on a phone?
In the arousal comparison, calmer content such as TV was associated with roughly 3.5 minutes more sleep than alerting content such as video games — a small difference. The larger difference is structural: an episode with autoplay disabled has an ending, and a feed does not.
Will a blue-light filter or night mode fix the problem?
It addresses the smallest of the three mechanisms. Screen light was associated with about 2.7 minutes of delay across 11 experimental studies, so a filter is working on minutes. The displacement of bedtime is where the tens of minutes are, and no filter changes that.
Why does the advice say one hour specifically?
No trial settled on one hour. The figure comes from expert guidance, and the National Sleep Foundation’s 2024 panel reached consensus on only 5 of its 10 statements about screens and sleep — and every agreed statement that named an age group named children and adolescents, not adults. One hour is a reasonable planning margin rather than a measured threshold.
I am in bed for the hour because my living room is cold. Is that a problem?
It works against you more than the screen does. Spending the hour awake in bed strengthens the association between the bed and alertness, which is why keeping the bed for sleep is a standard component of cognitive behavioural therapy for insomnia. A chair with a blanket is the cheaper fix.
Can I use the hour for a bath, or is that too close to bedtime?
Timing is the whole finding. The pooled result that showed roughly 10 minutes faster sleep onset used water at 40–42.5 °C finished 1–2 hours before bed, which puts the bath at the start of the wind-down rather than at its end.
If the hour is right and the bed still is not
A wind-down hour only settles the time you lie down. It says nothing about what happens afterwards, and the two problems are easy to confuse — a correct bedtime followed by twenty minutes of rearranging reads like a schedule failure when it is usually a loft failure.
If that is the pattern, the thing worth measuring is the gap between your ear and the mattress in your usual position, not the clock. Our own catalogue sits at the firm end: the Aeris Contour Pillow carries 12 cm and 10 cm ridges on opposite edges, so a single pillow offers two lofts. Figures below come from each maker’s production spec.
Advertisement
FirmNest Extra Firm Pillow
64 × 40 × 14 cm, 1.5 kg extra-firm memory-foam core.
See the FirmNest Extra Firm Pillow on AmazonProducts mentioned in this guide
Aeris Contour Pillow — 60 × 40 cm, 12 cm and 10 cm contour ridges, 1.4 kg memory foam
FirmNest Extra Firm Contour Pillow — 60 × 40 cm, 12 cm and 10 cm ridges, 1.4 kg, firmer core
