Med thumb sleep debt clock feature layout main

What is sleep debt? Strictly speaking, it’s the difference between the amount of sleep we get and the amount we should be getting, the idea being that the ill effects of missing sleep accumulate like mounting financial debt. We pay interest in the form of everything from health problems and reduced functioning to non-stop yawning.

As our Sleep Debt Clock shows, America’s sleep debt grows by 121 days with each passing second. On the weekends, the rate slows to a moderate 61 days-per-second. To arrive at these incomprehensibly large digits, we consulted sleep studies, health surveys and experts in the field. We identified both the amount of sleep people should get each night and the amount they do get. Then, we tallied up the difference between those numbers to determine their sleep debt, and created a formula to calculate the rate at which that debt grows.

Is it perfect? No — there’s no way to account fully for the nocturnal experiences or biology of 294,975,315 people. But you are in there somewhere.

Primarily, we see the Sleep Debt Clock as a way to highlight substantive discussions underlying the ticking number. And we welcome debate — about the public health impact of poor sleep, the quest to measure sleep debt and the merit of sleep debt as a concept in the first place.

For the sake of transparency — and because it’s compelling stuff — here’s a rundown of how we arrived at the debt-clock formula as well as an overview of the numbers we used, an explanation of how we chose them and some observations we made along the way.

The Formula

Final Clean Sleep Debt Formula

Some Key Terminology

Rate: On average, American adults sleep around seven hours a night. That number jumps on Friday night for everyone except the 55-plus crowd. Since most people get more sleep on weekends, the clock slows on Saturday and Sunday.

Population: The clock tallies our collective national sleep debt, which grows by the second. In this case, “national” is referring to American citizens between five years old and 85 years old, minus certain institutionalized populations for whom we couldn’t find reliable, meaningful data. (More on that below.) We broke down the general population into children, teens, adults and older adults to reflect shifting sleep needs and habits. We eliminated kids less than five years of age, both because we couldn’t find reliable numbers and because poor sleep isn’t yet a systemic problem for the toddling crowd.

Actual sleep: The average amount of sleep that members of a given population get each night. We found this information in the National Sleep Foundation’s yearly Sleep in America polls, which include nationally representative samples of between 1,000 and 2,000 people. In the NSF polls, participants reported the average amount of sleep they get on weekday and weekend nights. Their responses were recorded in hourly ranges: less than 6, 6 to 6.9, 7 to 7.9, 8.0 to 8.9, more than 9. From there, they calculated mean and median. We took the mean and compared their sleep duration averages against data from government surveys, for accuracy and perspective.

Ideal Sleep: This is the amount of sleep that members of a given population should get each night. We used the NSF sleep duration recommendations, adjusted for age group. We also consulted researchers and reviewed studies to choose numbers rooted in as much consensus as possible. The American Academy of Sleep Medicine (AASM) told us “anything below 7 hours is suboptimal.” We felt like “suboptimal” was a cut-off with a low bar. So, we went with NSF duration recommendations from 2015.
Note: Without adjusting for age, adults should shoot for a #hardeight. Older adults need slightly less, so we used 7.5 as ideal for over 55 and up. Kids and teens need more: 6-to-9 year olds need 10 hours and 12-to-17-year-olds need nine.

The Data We Used

Chart Image_DebtClock

It’s easy to find numbers to back up our national shuteye scarcity. Government agencies, non-for-profit organizations, polling services and academic research groups all gather and publish data that either specifically concerns American sleep habits or includes sleep-health metrics, but we chose NSF data for its breadth and ease of use. Here are some of its strengths and weaknesses.

National Sleep Foundation (NSF) Sleep in America Polls

NSFSleepinAmericaPollSleepDebtClockBackup

Pros: NSF data was abundant and easy to use. Since 2002, the foundation has published 13 annual Sleep in America polls, with different but overlapping topical focuses (e.g.,”sleep and the family,” “international sleep,” “adult sleep”) on populations across the age spectrum. Similar-enough questions about sleep duration appeared over the years, which let us check the NSF data against itself.  

Cons: Sample sizes, which range from 1,000 to 2,000, are small for a nationally representative survey of this kind. The 2015 Sleep in America poll included 1,029 people and focused on sleep and pain. Additionally, experts who weighed in favored data from government sources over the NSF because it has a range of corporate sponsors.

Vetting the Numbers

Since we used NSF sleep-duration data, we checked it against numbers from sources with more nonpartisan interests, and that’s exactly what we did. Here are some of the sources we used:

1. The Bureau of Labor Statistics (BLS) Annual Time Use Survey (ATUS)

AnnualTimeUseSurvey_Screenshot

Pros: It offers a hard number for average sleep duration: 8.4 hours.

Cons: Its duration figure is an outlier from the 7-ish average we consistently found elsewhere, and it’s not clear why.

Note: Last spring, we saw a spate of articles with headlines like “Americans actually do get enough sleep.” These stories all referenced the ATUS data. Some stories reported the results without much analysis; others argued that ATUS numbers are the unbiased, accurate estimates.

We reached out to the Bureau to make sense of their deviant data, but didn’t get much clarity. So we came up with our own hypothesis: The survey’s primary focus on time use affected how people tabulated sleep. Perhaps, they first estimated time spent on other activities (work, play, nihility) and delegated the leftover hours for sleep. In other surveys, respondents thought about sleep first, not as the final step of an elimination process. This is speculation, obviously. In general, we think the methodology for asking about sleep habits merits a closer look.

2. Gallup 2013 sleep poll

GallupPollSleepDebtScreeshot

Pros: The findings are clear and echo the seven-hour figure. The poll breaks down sleep data down by gender, age, employment status, household income and parenting status. Results are recorded as percentages of respondents getting sleep within hourly ranges. But approximated hour/minute conversions are available.

Cons: It only has a sample size of 1,031.

3. The Cleveland Adolescent Sleepiness Questionnaire

Clevelandsleepinesspoll_main

Pros: It offers in-depth insight on teen sleep habits and the authors actually calculated sleep debt: Per day, teen participants averaged 1.7 hours of sleep debt.

Cons: It only has a sample size of 411 and is not nationally representative.

4. Centers for Disease Control and Prevention (CDC) Surveys: Researchers uniformly steered us towards nationally representative and reliable reports from the CDC that include questions on sleep health. To the best of our ability, we checked NSF averages against these CDC surveys.

A) Behavioral Risk Factor Surveillance System (BRFSS)

Pros: Over 500,000 people participated in the last BRFSS. It’s thorough, reliable and clear.

Cons: The survey records sleep duration as either sufficient or insufficient, using seven hours as the cutoff, which wasn't terribly compatible with our quest for hard averages. For example, 35 percent of adults (18+) say they average less than seven hours of sleep per night.

B) National Health Interview Survey (NHIS)

Pros: The NHIS has a fairly large sample size — 112,000 people — and has asked about sleep for a few decades, which makes it useful in assessing how and how much American sleep habits have changed over time. For example, in a 2015 analysis of NHIS data, researchers found that average sleep duration has decreased slightly since 1985 and the percentage of adults getting six-or-fewer hours of sleep has increased by almost one-third. These numbers, however, haven’t changed much since 2004. Other CDC surveys, by contrast, only added sleep questions in the 2000s. Recent surveys are also useful in assessing sleep habits with respect to family structure.

Cons: Participants are between 18 and 64 and all live in single-family homes, which excludes anyone living with roommates or in other group-housing arrangements. Results, then, may not accurately represent considerable chunks of the country, including millennials, urban residents and lower-income Americans.

C) National Health and Nutrition Examination Survey (NHANES)

Attributes: “Adult” means people 16 years and older. It’s hard to find data from large surveys that reflects input from teens themselves rather than parent proxies, so answers from that two-year age group are valuable.

Shortcomings: With a sample size of 6,900 (for the most recent survey), it’s a bit smaller than the other CDC surveys. And, the data is hard to work with for anyone interested in hard numbers. Good for comparison.

Whose Debt is It Anyway?

Person Sleeping_Debtclock

We broke the general population into age groups, but we also wanted the formula to account for groups whose sleep habits either deviate from the general population considerably and meaningfully or who aren't represented in the national surveys.

* Deviant sleepers included shift-workers, people in low-income brackets, racial minorities, residents in atypically sleep-deprived regions and undocumented citizens.

* Unrepresented groups included military, the incarcerated, other institutionalized groups (e.g., psychiatric in-patients, nursing home residents).

We found plenty of data on the above-mentioned groups. In some cases, it wasn’t compatible with our formula and making it work was beyond the scope of our project. In other instances, we realized sleep duration wasn’t all that relevant to the group’s sleeping issues, and that their averages still resembled those of the general population. 

The incarcerated were a primary example of a population with widespread sleeping problems for whom we only found scarce, incompatible data. We could have used inmates’ lights-out/wake-time schedules as a proxy for sleep duration, but didn’t feel confident that internally imposed structure translated to hours snoozing. In fact, we’re confident it doesn’t, because we found evidence of under-recognized sleeping problems that we’ll be discussing in an upcoming series on sleeping in prison.

Of the above-listed groups, we only came up with an average sleep-duration number for the military: 6.5 hours per night.  

In either case, that data came from:

The Millennium Cohort Study

Backstory: We used the Millennium for military sleep. The study reflected data collected from 41,225 enrolled military who fought in Iraq or Afghanistan. They initially took the survey between 2001 and 2003, before they’d been deployed. Then, participants provided follow-up answers between 2004 and 2006, by which point their deployment statuses varied.

Methodology: Sleep duration averages weren’t hard to find, but they varied considerably across studies. In four studies concerning different military subsets, estimates ranged from 5.8 hours to 6.5 hours, without any discernible link between duration and deployment status.

Note: In this case, we chose the highest average, 6.5 hours, for a few reasons: The number emerged in the largest study, reappeared in other studies and across deployment groups in individual studies, was endorsed by a comprehensive and trustworthy review paper and, as was the case with our other numbers, reflected self-reported data rather than results from physiological sleep tests — apples to apples.

So, Can We Repay Our Debt?

Not really, but freezing it would be great. 

Sleep debt is a simple concept: We accrue it when we get less sleep than we should. So, by getting enough sleep, we could slow the rate of growth. But, most of our debt is a sunk cost.

For starters, sleeping too much isn’t healthy. We don't want to reverse the debt for the sake of doing so. And, from a scientific standpoint, it’s not clear that lost hours are fungible in terms of their health impact. For example, can you make up for a three-hour deficit on Tuesday by sleeping for 11 hours on Sunday? Numerically, sure. But, the recovery hours may not undo ill effects of lost sleep. That’s not to say catching up on sleep is bad for you (it’s lovely!), or that those extra hours don’t come with other benefits (better skin, better mood). Rather, it means that “dollar in, dollar out” doesn’t translate to shuteye. 

Based on research, however, we decided it does make sense to balance your shuteye budget on a week-to-week basis, if for no other reason than to keep sleep health on your mind.

Let’s work to freeze that clock. Or maybe, we should reset the clock and start fresh.