The ATUS is out! And yep, we sleep.

photo-128The Bureau of Labor Statistics released the annual American Time Use Survey results today. This is my favorite time study for a few reasons. First, it is huge (many thousands of people). Second, a few features help with accuracy. A researcher asks the participant to go through yesterday, which means the participant isn’t trying to recall a typical day. There are no typical days. Since the survey isn’t targeting any particular category of time, people are less inclined to give socially desirable answers. This methodology produces some interesting findings:

The average amount of sleep in 24 hours is 8.74 hours. That’s 8.65 hours for men and 8.82 hours for women, and is 8.48 hours for weekdays, and 9.34 hours for weekends/holidays.

Of course, that’s averaged over all Americans aged 15 and older. If we look at those with kids under age 6, we get 8.6 hours/day (men = 8.19, women = 8.93). If we look at those who are employed and have kids under age 6, we get 8.26 hours/day (men = 8.10, women = 8.47). For those with kids aged 6-17 we get 8.61 hours (men = 8.56, women = 8.65), and for those who are employed with kids aged 6-17 we get 8.38 hours (men = 8.36, women = 8.4).

Looking at the whole population, the average American spends 5.26 hours per day engaging in leisure and sports, which includes 2.77 hours watching TV as a primary activity. By contrast, the average American spends a mere 3.46 hours in work and work-related activities.

That’s partly a function of many people not being in the workforce, though. Among employed Americans, leisure drops to 3.73 hours a day. In its press release, the ATUS notes that “employed adults living in households with no children under age 18 engaged in leisure activities for 4.5 hours, about an hour more than employed adults living with a child under age 6.” The good news is that this lower number — 3.44 hours/day — is still pretty good. That’s about 24 hours per week. People who are employed and have kids aged 6-17 clock 3.95 hours of leisure activities per day, or a bit under 28 hours per week.

Where do you think you come out on the leisure total?

16 thoughts on “The ATUS is out! And yep, we sleep.

  1. What’s considered “leisure” here? As a single mom, I have about 1 hr per day that I am not working/sleeping/actively engaged in childcare, and even that’s sketchy now that my toddler is going to bed around 10:15!!!

    1. @gwinne – here are some selected characteristics: http://www.bls.gov/news.release/atus.t11.htm — TV, reading, socializing, exercising, etc.

      As a single mom with a full-time job, and two kids, including one very young one, you have several factors that would put downward pressure on the total. The kid sleep thing is a huge variable. Moms of young kids who go to sleep by 7:30 can get quite a bit of leisure time. Those whose kids don’t do that regularly face a different picture.

      1. Interesting and very slippery… I need to distinguish between reading for “leisure” and reading for work which is messy for someone who teaches literature! Likewise, while I’d count staring into space on my porch as leisure if that were my goal…but thinking/reflecting is a critical part of my job as writer/scholar/teacher!

        (I am counting writing THIS as leisure, but it’s a few minute break between writing tasks.)

        Have you read Schulte’s Overwhelmed? Makes an argument that the leisure time of working moms is very fragmented (10 minutes here and there and therefore potentially less restorative…)

  2. I got a FitBit One earlier this year, and one thing that I noticed right away was that I was overestimating the amount that I slept the night before. If I went up to bed at 9pm and then got up at 5am, I would count that as 8 hours of sleep. However, what would really happen is that I’d go up to bed, do all the “getting ready for bed” stuff, read some, and then go to sleep, and my sleep would be interrupted by baby wakings and bathroom trips. In all, I was getting 1/2 to 1 hour less sleep than I had estimated (and meant I needed to go upstairs earlier). That doesn’t exactly have anything to do with the sleep study, but between that and the other FitBit talk, it popped into my mind.

    As a side note, it might be worth presenting the median as well as the mean time for these things. I’m always suspicious of means when I haven’t yet been convinced that data is normally distributed. Sleep I could believe, but I’m not sure about leisure time. And, because of the unemployment issues you mention, I don’t think it’s very meaningful to talk about the average number hours worked. Again, with all those 0s, it’s not going to be normally distributed.

  3. Guess it looks like that’s all they provide, though. Too bad! For some reason I thought the released the raw data and you did your own analyses.

    1. @Chelsea – yes, median and mean can make a difference. In this case, though, we’re working within a relatively narrow range: 0-24 hours, and some of the big potential zero items, like work hours, are broken out by employed, not employed. No one can spend 10,000 hours on leisure activities in a day, unlike in a survey of, say, net worth, which could be all over the place.

    2. The raw data for 2002-2012 are available for download either directly from them or from a super easy to use (if you do data analysis) site that they link to where they include the CPS variables on these participants too. I assume the 2013 haven’t been cleaned and added to the easy to use site yet. The medians for the 2011-2012 are a bit lower than the mean– but a lot of that is probably heaping at 8.5 hours.

      1. I do a lot of data analysis, but not with survey data so I’ve been leery of digging in. Do you need to use something like Stata to incorporate a weighting scheme due to the survey design into your models? Or can you use it as is and assume that, because of the survey design, it is representative of the US population?

  4. When I do time logs (I’ve done 3 1-week blocks), I count “sleep” as the time I go up to go to bed to the time I get out of bed. Because I”m not doing anything else in those times. But there are often full hours on either end of the night that I am not yet asleep (I go up, lay down, almost fall asleep, husband comes up and turns on light, I’m up again for 30 minutes) or am awake and not yet up (wake up, realize its only 5:30, try to sleep more, fall asleep at 6:20 and then get up at 6:30 when son wakes up). That doesn’t even count the night wakings (son wakes up at 1:30, I go into his room and soothe, back to bed, wakes back up before I fall back sleep at 1:50, I bring him into bed. He squirms and paws at me for 30 minutes and falls asleep, I fall asleep 30 minutes later).
    Sure I have 8 hours devoted to “sleep”, but maybe slept 5 fragmented hours during that time.
    What i mean is, its hard to really quantify, and if people feel they don’t get enough sleep, that they are tired, well they probably are.

    1. Like Gwinne, I have the horrible vicious cycle combination of poor-sleeping kid and insomnia (trouble falling and staying asleep, including early early morning waking) so I guess I’m a bit prickly about the subject!

  5. Ana and Gwinne, what a wonderful thing to realize. The question now is “What are you going to do about it?”. When my 2 children were toddlers plus, I set out to fix their sleep patterns to improve my own. Changes or identifying things like no caffeine after 2 pm, playing with them for an hour after dinner, and keeping them on an evening routine resulting in improved quality time with my kids and quality sleep time for all of us. Good Luck!

Leave a Reply

Your email address will not be published. Required fields are marked *