In the field of time use, it’s one of the great questions. Everyone thinks Americans are increasingly sleep deprived. Giant surveys such as those done by the National Sleep Foundation often provide evidence of this. And yet the American Time Use Survey consistently shows that Americans — even those who work and have young kids — sleep 8-plus hours per day.
Why the difference? A key issue is that the NSF’s annual survey has historically asked about a typical night (see the 2008 summary of findings, p. 10 shows charts for a “typical” night). We may build our lives in ways that don’t seem to allow for enough sleep in a normal day’s schedule. Yet we may sleep at unscheduled times. That sleep gets counted in a time diary, and is at least one of the reasons that the totals are different.
I was thinking of this during this past weekend, when I completely conked out on Saturday afternoon. I’d probably been averaging about 7 hours of sleep a night over the previous few days. Not bad, really. But when my 2-year-old fell asleep at 2 p.m. on Saturday, I lay on my bed to read a novel, and promptly fell asleep myself. I woke up at 4:40 p.m. when my husband put the (awake and whining for mommy) toddler on top of me.
I’m not suggesting that most people have that experience. But this extra sleep can affect an average a great deal. If I slept 7 hours a night for 7 days, but then added in an extra 160 minutes of napping at some point, that gets me to a daily average (calculated weekly) of about 7:23.
That’s not drastically different, I suppose, but there turn out to be lots of occasions for unscheduled sleep. People set their alarms, and then hit snooze three times. Getting an additional 27 minutes of sleep on one morning, averaged over 7 days, adds 4 minutes per day to the total. And people who hit snooze 3 times don’t generally just do this one day per week. Falling asleep at 9 p.m. in front of the TV one night, instead of the “typical” 11 p.m. adds 17 minutes to the daily average. Do it twice and you’ve added more than half an hour. Sleeping in on weekends, of course, can up a daily average (though the National Sleep Foundation does distinguish between weekday nights and weekend nights, and the ATUS too — and in both cases the ATUS totals are higher).
There’s also the matter of what we consider “typical.” Not all weekdays may be exactly the same, but if we’re feeling particularly tired, we may remember the one that involved staying up until midnight working as typical, instead of the good ones. All this contributes to different totals in our minds, vs. what would actually be recorded in a diary.
Of course, the question of whether bad nights are mitigated by unscheduled sleep isn’t a completely settled one. Runners who get a bad night of sleep right before a race tend to do OK if they’ve had good nights prior. On the other hand, most sleep doctors say you can’t completely make up a deficit on weekends. From seeing hundreds of time logs, my take-away is that each person tends to require a certain amount of sleep, and one way or another most of us who aren’t tending to 2-week-old babies get it. If you need 7 hours, and your “typical” night only accounts for 6, you’ll fall asleep in front of the TV on Wednesday night and oversleep your alarm on Friday. That late arrival at work wasn’t part of the schedule, but the sleep still happened.
Do you have unscheduled sleep in your life?
Unscheduled sleep? Not without a head injury or major illness.
I’m a major insomniac with a toddler who doesn’t sleep through the night. I slept a fragmented 4.5-5 hrs last night, which is literally half of what I need to feel rested.
I’m one person, to be sure, but MANY Americans have significant insomnia and sleep disorders. And 7 hours plus a short nap is NOT qualitatively the same as 7.5 hrs sleep…
@gwinne- an average doesn’t mean anything about one individual person. There are no doubt many people who are sleep deprived, have insomnia, etc., and an average shouldn’t be interpreted to mean that your experience (or anyone’s experience) isn’t real. But with the average coming well over 8 hours a night, that also means that many people don’t have that experience, including many parents of young children.
No doubt, as my original comment indicates. But what is average here (mean, median, mode) and how determined?
Also think it’s problematic to just go on number alone, as people have different sleep needs and there are different qualities of sleep.
@gwinne- on something like sleep I’d imagine that the median and mean wouldn’t be hugely far off, because it’s a limited range (0-24 hours per every 24 hour period). It’s not like wealth, where 1 person worth $10 billion mixed in with 10,000 people worth nothing makes everyone a millionaire.
I picked sleep for my example ATUS problem last semester (because it’s so important).
The mean for sleep (for everyone) is: 531.8292, the median is 524. The mode isn’t particularly informative because we’re in minutes and it looks like there’s a lot of heaping in the dataset, but it is 540.
Though the point Gwinne is making is actually about quantiles, not mean/median/mode. Running those:
1% = 211
5% = 330
10% = 380
25% = 450
50% = 524
75% = 600
90% = 693
95% = 761
99% = 930
So a wide range.
@Gwinne- these are minutes per day of sleep, looking at the whole distribution. So someone getting 5h30m (330 minutes) is in the 5th percentile of sleep. 95 percent sleep more. Someone who got 3h31m (211 minutes) is in the 1% — 99% of people slept more.
NandM…explain?
1% means that 1% of the people are 211 min or below for sleep. 5% means that 5% of the people are 330 or below for sleep, and so on.
*
The median is the quantile (percentile) that’s at 50%.
Your distribution makes me feel like much less of a freak for getting 8.5h (510 min) per day. Thanks for this!
Sleep is a mess around here lately. Part of it is the construction so we’re all sleeping in the living room, and 3 months of sleeping on the couch is taking its toll. But the other part is that my 15 month old is highly unpredictable – she’ll wake anywhere from 0 to 3 times a night.
I’ve made a conscious effort to prioritize sleep because I’m so annoyed by moving through life in a tired haze. I’ve been aiming to get to bed before 11, but I’ve also been falling asleep during the kids’ naptime/quiet time, which sometimes makes things worse at night.
In my ideal world I would not nap, but get to bed early enough to get about 9 hours. That two hour block mid-day while the kids nap would be prime work time for me if I could just stay awake and focused.
Thanks for this! My husband falls asleep on the couch pretty much every single night. He says that he actually LIKES doing this, even though I’ve told him repeatedly that his sleep quality will be better if he just goes to bed. Anyway. Lately he’s been going to bed at 10 and waking up at 4:40, but with the extra couch sleep you can probably add 30 minutes or so of sleep each day, which brings his total up to not terrible.
I will say that I wonder how I’m going to survive residency. With one day off per week and 13-14 hour days I suspect there won’t be much time available for “catching up.” And then there’s all the night shifts too. Ah well.
@oldmdgirl – I don’t like falling asleep on the couch either. I prefer my bed. When I’m feeling sleepy, I try to head there! Not sure what to say about residency.
My husband does this as well, I just head up when I feel tired! I can’t nap, never have (since about 12 months old, if you ask my poor mother), so its difficult for me to make up sleep. When I get my bouts of insomnia I’m really miserable, so I try make up for it by getting adequate rest when I don’t have insomnia.
As for residency, I know things in terms of hours and schedules are different now, but 1) I was definitely chronically sleep-deprived and 2) I am glad I didn’t have kids during my training because I would sleep whenever I could post-call (at least when we started being allowed to leave by 1pm the post-call day rather than working all day). When I had rotations that required early AM starts (i.e. PICU when we had to be there at 5) I would go to bed as soon as the sun went down (it was summer, so maybe 8pm?) but I only had myself to worry about it which made is possible.
Totally planning on going to bed at 8 or 9 most of the time for the next 4 years. Thankfully the offspring goes to bed then too and is a good (knock on wood) sleeper. Maybe I’ll be better rested during residency than those other childless youths because the offspring has instilled this additional level of discipline on me. Haha.
My problem isn’t thinking I sleep less than I do. I go to bed consistently at 10 and wake with my children usually between 6:15-7.
My issue is that I wish i did not require so much sleep, as even sleeping 8+ hours I feel tired.
@Mary- oh, I know the feeling. I don’t need that much sleep (I like 7.5 hours and a bit less is totally doable if I’m not exercising too intently), but I’ve often thought how much more me time I’d have if I only needed like 6.5 hours/night. But if I’m already at the 25th percentile, hoping to be in the 10th is probably just getting greedy…
OK I can’t reply directly to nandm’s data above, but WHOA the 95th percentile is >12 hours of asleep! so 5% of the respondants reported an average of >12 hours of sleep per day? 25% reported 10+ hours a day. Can this be real data, since 7-9 hours is the general amount most adults need. I feel like I sleep way more than most, and I get 7-8 hours per night.
The ATUS isn’t a long-term average of how much people are sleeping, just a snapshot. Those 5% of people could have the flu, or could be recovering from a week of being on-call. The universe also includes teenagers and older folks (age 15-85).
Or they could totally be like Brian Wilson. http://youtu.be/Ch84fmOa414
And you can see that the middle 50% (that’s 25-75%) are getting between 7.5 and 10 hours/night. That’s half of people.
And that’s a HUGE range. Wow. (And thanks for the new vocab word!)
*
But really–and this goes partly to my original question about data collection and averages–this isn’t EVERYONE, just the people tested, their faulty perceptions of their own sleep, etc.
But I think Laura’s question about perception vs what appears like “enough” sleep might have to do with untestable things. Like, if I get 7 hrs of sleep, and they’re the “wrong” 7 hours (say 9pm -4 am) I’m more tired than I am if I sleep 6.5 at a better time (like 12-6.30 am). And none of that is the same as getting as much sleep as I NEED, which is about 8.5… Talking about adequate sleep is not as simple as doing math and adding up hours…
If I limit to 25-55 year olds, the top tail drops off a bit, though the bottom tail is about the same.
1% = 210
5% = 330
10% = 374
25% = 445
50% = 510
75% = 600
90% = 685
95% = 750
99% = 908
A nurse who works in a nursing home commented that people over 75-80 generally sleep a LOT. I wonder if these numbers include elderly people at the high end.