Self Lab

Evidence-based self-improvement. No mechanism, no article.

Esc or click outside to close

Sleep & Energy · Mechanism

The Snooze Button Myth: It Costs 6 Minutes, Not Your REM

A sleep lab wired up 31 habitual snoozers: three snoozes cost about 6 minutes of sleep, not your REM. The real lever isn't the button, it's your bedtime.

Alex by Alex July 29, 2026 17 min read

The alarm goes off. Your hand finds the snooze button before your eyes open. Nine minutes later, again. And somewhere under the fog sits a small guilt, because you’ve read that you’re shredding your REM sleep, spiking your stress hormones, and starting a sleep cycle you can never finish.

Almost none of that has been measured. What has been measured is stranger and more useful.

A Stockholm University lab put 31 habitual snoozers through three nights of sleep monitoring and compared snooze mornings against straight-through mornings (Sundelin et al., 2024). Across the whole night, nothing changed. The total cost was about six minutes of sleep.

This isn’t permission to snooze forever. It’s a swap: stop fighting the last alarm, and start moving the thing that actually predicts it. Because the button is a symptom, and motivation at 6am is a mood, not a fuel tank.

The Bottom Line

  • Three snoozes across 30 minutes left whole-night sleep statistically unchanged in the lab (Sundelin et al., 2024).
  • The real price was roughly 6 minutes of sleep, and that half hour still held about 23 minutes of real sleep.
  • The REM and cortisol scares aren’t in the data. Neither is a proven benefit.
  • Snoozing tracks being an evening type and how last night’s bedtime moved against your own usual, not laziness.
  • The lever worth pulling is a steadier bedtime plus a 20-minute buffer after waking.

A quick note: this is education, not medical advice. The lab study below deliberately excluded people with insomnia symptoms, poor sleep quality, or signs of sleep apnea. If your mornings are wrecked by a sleep problem, none of this applies to you. See a clinician instead.

In this guide:

The real lever sits the night before, and an afternoon coffee reaches further into it than most people think.

What does hitting snooze actually cost you?

About six minutes of sleep. Not six minutes an hour, and not a ruined night. Six minutes, total.

Here’s the setup. Sundelin and colleagues published the experiment in the Journal of Sleep Research in 2024 (Sundelin et al., 2024). Thirty-one habitual snoozers slept in a lab with full sleep monitoring on their heads. Each person did both conditions: one morning with three snoozes spread over 30 minutes at 9 to 10 minute gaps, one morning with a single alarm. Same final wake time. Same time in bed.

Across the whole night, no sleep measure differed significantly. Not total sleep, not deep sleep, not REM, not how often they were aroused.

What was measuredSnoozedSlept straight through
Total sleep, whole night418.0 min416.6 min (p = 0.805)
Deep sleep (N3), whole night98.0 min95.4 min (p = 0.617)
REM, whole night84.7 min90.1 min (p = 0.263)
Arousals per hour, whole night5.65.6 (p = 0.894)
Sleep in the final 30 minutes23.0 min (median)29.0 min (p < 0.001)

Source: Sundelin et al., Journal of Sleep Research, 2024.

Look at that last row again. The snooze window did cost real sleep, and the loss was statistically solid. But it still contained roughly 23 minutes of actual sleep. Snoozing is not half an hour of lying there awake, punishing yourself. Sleep efficiency in that window dropped from 95.1% to 75.4%, which sounds dramatic until you translate it: you slept three quarters of the time instead of nearly all of it.

Straight talk: every person in this study snoozed at least twice a week and fell back asleep between alarms almost every time. The researchers screened out insomnia, poor sleepers, and breathing problems. If you lie awake staring at the ceiling between alarms, you are not the person this was measured on.

Does snoozing wreck your REM sleep?

No. The most repeated version of the myth is the one the data contradicts most directly.

You’ll find the claim on major health sites: the last stretch of the night is mostly REM, snoozing interrupts it, and that interruption triggers a fight-or-flight response that raises your heart rate and blood pressure. Cleveland Clinic’s explainer says exactly this, and cites no study for either half (Cleveland Clinic, 2025).

Now the measurement. In the snooze condition, the final 30 minutes contained a median of zero minutes of REM. There was almost nothing there to interrupt. Waking out of REM was actually rarer when people snoozed, happening to 6 of them versus 10 in the straight-through condition. Whole-night REM didn’t differ (p = 0.263).

The cortisol scare doesn’t survive contact with measurement either. The researchers collected saliva and found a normal cortisol wake-up curve in both conditions, with no main effect of snoozing. One exploratory test found slightly higher cortisol at the moment of waking after snoozing, and the authors themselves say more data is needed before anyone reads anything into it. Forty minutes later, no difference at all. Sleepiness ratings and mood were flat between conditions too, and there the statistical support was strong.

The “you start a new sleep cycle you can’t finish” line has the same problem. In the measured snooze window, nobody entered deep sleep at all. It was light sleep, the shallow stuff, which is the whole proposed point.

Here’s the mechanism the authors suggest, and it isn’t extra rest. It’s stage management. The first alarm lifts you out of deep sleep early, the snooze window parks you in light sleep, and the final alarm has less distance to drag you. Treat that as a plausible explanation supported by the stage data, not a proven chain.

So is snoozing good for you?

Also no. And this is the part almost every write-up got wrong in the other direction.

Right after getting up, the snoozers took four short cognitive tests. They did no worse on any of them. On three, they did slightly better: arithmetic speed, memory for things they’d learned, and the cost of mental conflict. That sounds like a win, and headlines treated it as one.

Then the authors did something the coverage skipped. They reported that the statistical evidence for nearly all of those differences is too thin to trust. Their words: the Bayes factors suggest that more data are needed in order to draw reliable conclusions for virtually all of the above effects. They also stated the tests were exploratory, weren’t corrected for running many comparisons, and that no power analysis was done because the sample size came down to available resources.

Scientific American’s write-up reported that the benefits lasted 40 minutes after rising (Solis-Moreira, 2023). Only one of those advantages was tested against time and lost it: the memory benefit was gone by the 40-minute retest. For the other two the paper found no interaction with time, which its own discussion says suggests the advantage held across the 40 minutes. What every write-up skipped is the authors’ verdict that the Bayes factors are too thin for virtually all of these effects.

So both camps overstated the same 31-person study, in opposite directions, using the same numbers. The anti-snooze coverage invented harms that weren’t measured. The pro-snooze coverage promoted a weak hint into a benefit and left out the authors’ own caveat.

The honest summary is duller and more trustworthy. In habitual snoozers, under one specific protocol, snoozing did no measurable damage overnight or that morning; the one debit the study did find was later in the day, with snoozers rating themselves slightly sleepier at lunchtime and reporting more effort on afternoon tests. A benefit may be there. The evidence isn’t strong enough to promise it.

Why do two sleep labs disagree?

Because they didn’t test the same behavior. One study found no harm. Another found snoozing made grogginess worse. The difference sits in the gap between alarms.

In 2022, Ogawa and colleagues published a lab study in the Journal of Physiological Anthropology with 10 habitual snoozers (Ogawa et al., 2022). By the third test block, roughly 22 minutes after waking, snoozers were slower: 238.4 milliseconds versus 221.9 (p = 0.024). The first two blocks showed no difference, which is what “prolonged” inertia looks like. Their sense of vigor rose after waking in the single-alarm condition, but not after snoozing. Their conclusion: snooze alarms prolong grogginess, possibly because they force repeated awakenings.

Sundelin et al., 2024Ogawa et al., 2022
People31 habitual snoozers10 habitual snoozers
Snooze window30 minutes20 minutes
Gap between alarms9 to 10 minutes5 minutes
Alarms before getting up3 snoozes4 alarms
Fell back asleep between alarmsrequired at screeningmeasured: 15.7 of the 20 min still asleep, but 3.9 min awake and 6.1 min in N1
Resultnight unchanged, small early edgeslower reactions, longer grogginess

Note what the 5-minute gaps did and didn’t do. Ogawa’s snoozers still slept 15.7 of those 20 minutes, so the window wasn’t sleepless. It was fragmented: 3.9 minutes awake against 0.2 in the single-alarm condition, 6.1 minutes of light N1 against 1.5, and 12.2 stage transitions against 3.5. The two labs also differ in who they tested, 10 students against 31 people screened for reliably falling back asleep, and in what they measured, an auditory reaction task about 22 minutes after waking. That’s a hypothesis for why the results split, not a finding either paper proves.

Keep the scale in mind. One study had 31 people, the other had 10. Neither is definitive, and anyone selling you certainty here is selling.

A third dataset sits in between. Researchers analyzed 103,169 nights of wearable data from 450 people (Mattingly et al., 2022). Snoozers didn’t sleep less: about 8.15 hours versus 8.21. In the final hour they logged 2.19% more light sleep and 1.86% less deep sleep. Snoozers also ran a resting heart rate about 1.86 beats per minute higher across the night. Read that carefully: it’s a trait difference between people who snooze and people who don’t, not evidence that snoozing raises your heart rate. The authors also flag that wearable sleep staging isn’t lab-validated.

Is sleep inertia caused by snoozing?

No, and this is the quiet second myth. Sleep inertia, the groggy stupid stretch after waking, happens to everyone, every morning, snooze or no snooze.

How long it lasts depends on what you measure. The Sleep Research Society’s public guide says symptoms typically fade within about 20 minutes (Hilditch & Vincent, 2023). A 1999 lab study that tracked alertness and mental throughput for four hours after waking found something slower: inertia faded gradually and took two to four hours to level off, with time constants of 0.67 hours for alertness and 1.17 hours for performance (Jewett et al., 1999).

Different measures, different literatures, different answers. Feeling normal comes back fast. Performing normally takes considerably longer.

And here’s the awkward bit for the story everyone repeats. In that 1999 study, the sleep stage people woke from made no significant difference to how severe their inertia was or how fast it faded. Neither did showering, eating, or turning on the lights. The Sleep Research Society’s guide does list waking from deep sleep as an aggravator, so this is not settled. But “you’re groggy because your alarm caught you in deep sleep” is a tidier explanation than the evidence has earned.

Which reframes the whole argument. If grogginess is standard equipment and takes real time to clear, then the interesting question isn’t whether to snooze. It’s what you schedule inside that first messy stretch of your day.

Who hits snooze, and why?

Most people, and not for the reason you’d guess. Two datasets that share no methodology land in roughly the same place.

A survey of 1,732 people found 69% snooze at least sometimes, spending an average of 22 minutes at it, with about 8 minutes between alarms (Sundelin et al., 2024). Separately, researchers pulled telemetry from 3,017,276 sleep sessions belonging to 21,222 phone-alarm users across four continents (Robbins et al., 2025). 55.6% of those sessions ended with the snooze button, averaging 2.4 presses and 10.8 minutes.

Self-report and machine logs rarely agree this well. When they do, the behavior is just normal.

Now the reasons people gave, which is where the character-flaw framing falls apart.

Why people say they snooze In the survey reported by Sundelin and colleagues in 2024, the most common reason for snoozing was being unable to wake up or too tired at about 25 percent. It feels good and waking up more gently each drew about 17 percent. Worry about not waking drew 9.5 percent, wanting more sleep 6.8 percent, and laziness only 5.5 percent. Why people say they snooze Share of snoozers giving each reason, survey of 1,732 people Can't wake up / too tired 25% It feels good 17% To wake up more gently 17% Worried I won't wake up 9.5% To get more sleep 6.8% Laziness 5.5% 0% 10% 20% 30% Source: Sundelin et al. (2024), Journal of Sleep Research
One snoozer in eighteen blamed laziness. One in four simply couldn't get up. Source: Sundelin et al. (2024).

The profile of a snoozer is a body-clock profile, not a character profile. Snoozers were on average 6 years younger than non-snoozers. They were almost four times more likely to be evening types (odds ratio 3.7) and three times more likely to report waking up mentally foggy (odds ratio 3.0). They slept 13 minutes less on workdays, but no less on days off, and they rated their sleep quality the same.

Read that last part again. The gap only shows up on days when an alarm decides the ending. That’s a scheduling mismatch, not a discipline failure. Which is worth knowing before you add another item to the pile of things you blame yourself for.

What to change instead of the button

Move your bedtime and your first 20 minutes. Those are levers. Willpower at 6am isn’t.

In the telemetry, going to bed earlier than your own usual predicted fewer snoozes; going to bed later split people into two camps, skipping the button entirely or hitting it seven-plus times (Robbins et al., 2025). Not earlier than some ideal bedtime. Earlier than yours. The button is downstream of last night.

And sleep regularity has hard-outcome evidence behind it: across 60,977 people wearing accelerometers, the four more-regular fifths of sleepers had 20% to 48% lower all-cause mortality than the least regular fifth, and regularity predicted mortality better than sleep duration did (Windred et al., 2024). That’s observational, and the authors say clearly that regularity may be as much a marker as a cause. To be blunt about the thing this article will not claim: nothing here says snoozing shortens your life. It doesn’t.

If your bedtime slides late because your phone eats the last hour, that’s the actual problem to solve, and it runs on a mood loop, not a time problem.

What to do with the groggy 20 minutes

The Sleep Research Society’s guide lists countermeasures with lab evidence behind them (Hilditch & Vincent, 2023):

Their honest limitation, stated in the guide: these strategies were mostly studied in labs, and it’s unknown how well they work at home or at work.

So stop negotiating with the last alarm. Move the first one instead. Set it 20 to 25 minutes before you need to be functional, and put nothing that requires a decision inside that window. You aren’t trying to feel sharp at the moment you stand up. Nobody does.

FAQ

Is hitting snooze bad for you?

No measurable damage overnight or that morning. Three snoozes across 30 minutes left whole-night sleep, cortisol, mood, and morning sleepiness unchanged in a 31-person lab study, at a cost of about six minutes of sleep (Sundelin et al., 2024). The one debit the study did find was later in the day, with snoozers rating themselves slightly sleepier at lunchtime and reporting more effort on afternoon tests. Neither harm nor benefit is well established.

Does snoozing disrupt REM sleep?

The measured snooze window contained a median of zero minutes of REM, so there was essentially nothing to disrupt. Whole-night REM didn’t differ significantly, and waking out of REM was less common among snoozers, not more (Sundelin et al., 2024). The claim circulates widely without a study attached.

How many people hit snooze?

Most of them. 55.6% of 3,017,276 tracked sleep sessions ended with a snooze, averaging 2.4 presses and 10.8 minutes (Robbins et al., 2025). A separate survey of 1,732 people put it at 69% who snooze at least sometimes. Two independent methods, one conclusion: it’s normal.

Should I move my alarm across the room?

Maybe, but understand what it does. It forces one hard wake-up instead of a gradual one. In the lab study, nobody woke from deep sleep after snoozing while 30% (10/31) did when they slept straight through, and the authors suggest waking from lighter sleep may ease inertia (Sundelin et al., 2024). That’s a plausible mechanism, not a proven one. If you consistently oversleep and miss things, the friction still helps. If your problem is grogginess, skipping the snooze buffer might make it worse, not better.

What if I lie awake between alarms?

Then this research doesn’t describe you. Every participant in the lab study fell back asleep between alarms almost every time, and people with insomnia symptoms or poor sleep quality were screened out. Lying awake in bed for 30 minutes each morning is a different situation. A clinician is the right stop, not an article. Building the evening habit first helps more anyway, and there’s a way to size that habit so it survives bad days.

The Bottom Line

The snooze button is the least interesting part of your morning. It costs about six minutes of sleep, it doesn’t touch your REM, it doesn’t disturb your cortisol wake-up curve, and the case that it helps you is too thin to lean on. Meanwhile, grogginess after waking is universal, takes real time to clear, and doesn’t much care which alarm woke you.

What does move is the timing on either side. A bedtime closer to your own usual, and a buffer between opening your eyes and doing anything that matters. One of those you set tonight. The other you set right now, on your phone.

One small action today: open your alarm app and set your first alarm 20 minutes earlier than the moment you actually need to be sharp. Then move your first real task, the email, the drive, the decision, to after that 20-minute window. Change nothing else this week.

Alex is the voice of Self Lab: practical psychology for people who are done with motivational fluff.

Sources