The Sleep Revolution

For decades, sleep deprivation was America’s unofficial badge of honor. Margaret Thatcher reportedly slept four hours a night. Thomas Edison called sleep “a criminal waste of time.” By the early 2000s, the Gallup poll found 40% of Americans sleeping six hours or less — a fact that prompted an industry of books, TED talks, and wellness retreats dedicated to the gospel of more rest. Arianna Huffington collapsed from exhaustion at her desk in 2007 and turned it into a bestselling book. Matthew Walker’s Why We Sleep became a phenomenon in 2017, arguing that sleep deprivation was linked to everything from Alzheimer’s to car accidents to reduced sperm count.

But the data tells a story the culture hasn’t caught up with. According to the American Time Use Survey, the average American in 2003 slept 8 hours and 34 minutes per day. By 2024, that figure had risen to 9 hours and 2 minutes. That’s a gain of 28 minutes per day — equivalent to an extra 170 hours per year, or more than a full week of additional sleep annually. The increase wasn’t sudden. It accumulated slowly, about 1.3 minutes per year, a drift so gradual that it escaped the headlines entirely.

The gain is robust. It appears in every demographic subgroup the BLS tracks: men and women, employed and not employed, young and old, weekdays and weekends. It isn’t an artifact of changing population composition or survey methodology — the ATUS has used essentially the same 24-hour diary method since 2003. Something real happened. Americans chose, or were compelled by changing circumstances, to spend more of their fixed 24-hour budget in bed.

What’s striking is that the gain accelerated. From 2003 to 2010 — a period that included the Great Recession — the national average rose from 8.57 to 8.67 hours, a modest 6-minute gain over seven years. From 2010 to 2019, it climbed to 8.84 hours, adding another 10 minutes. Then came the pandemic. The 2020 survey was suspended entirely. When data collection resumed in 2021, the average had jumped to 8.95 — and by 2023 it peaked at 9.07 before settling to 9.04 in 2024. The post-pandemic gains did not reverse. America sleeps like a different country now.

The sleep revolution was not distributed equally. Every age group gained, but the biggest winners were not who you’d expect. The cohort that added the most sleep was ages 45 to 54 — the peak of career responsibility, the years of managing teenagers and aging parents and big mortgages. In 2003, this group slept just 8.17 hours, the least of any age bracket. By 2024 they were at 8.76 hours, a gain of 35 minutes per day. Whatever was keeping middle-aged Americans up at night in 2003, it has loosened its grip.

Young Americans, ages 15 to 24, gained 31 minutes — from 9.15 to 9.66 hours. But they started from such a high baseline that the question is less “why are they sleeping more” than “why did they already sleep so much?” This is the age group still in school, still living at home, still working part-time or not at all. A 15-year-old and a 50-year-old executive inhabit entirely different time economies. The 55–64 age bracket gained 32 minutes (8.36 to 8.89), tracking the national average. And retirees 65 and older — who already slept 8.86 hours in 2003 — gained a more modest 17 minutes, arriving at 9.15. When you’re already sleeping nearly as much as you want, the ceiling is lower.

The one group that stands out for its restraint is ages 35 to 44. They gained just 21 minutes, less than any other cohort. In 2024 they sleep 8.69 hours, the second-least of any age group. This is peak parenthood. Children under six live in their house. School drop-offs start at 7:30. The morning alarm is non-negotiable. These Americans are caught in the vice of biological and logistical necessity — and the data shows it.

Women sleep more than men. They have for every year the survey has been conducted, and the gap is remarkably stable. In 2003, women averaged 8.65 hours to men’s 8.48 — a difference of 10 minutes. In 2024, it was 9.11 versus 8.96 — a gap of 9 minutes. Both sexes gained almost exactly the same amount of additional sleep: women added 28 minutes, men added 29. The revolution was gender-blind.

The reasons women sleep slightly more are the subject of medical debate. Some researchers point to hormonal differences, particularly the role of estrogen and progesterone in sleep regulation. Others note that women are more likely to be out of the labor force — and as we’ll see, employment status is the single strongest predictor of how much someone sleeps. Still others cite higher rates of chronic fatigue and fibromyalgia among women, conditions that increase time in bed even if they don’t always produce restful sleep. The ATUS measures time in bed, not sleep quality — a distinction worth remembering.

What the data does show clearly is convergence on the margins. The gender gap shrank from 10 minutes to 9 — not dramatic, but consistent with the broader story: whoever was sleeping the least in 2003 gained the most by 2024. Men, who started lower, caught up slightly. The same pattern shows up in employment, age, and weekday–weekend splits. The sleep revolution has been quietly equalizing.

Nothing predicts your sleep like whether you have a job. In 2024, employed Americans slept 8.80 hours. Those not employed slept 9.43 hours — a gap of 38 minutes. That gap has actually narrowed since 2003, when employed workers slept 8.26 hours versus 9.12 for the jobless — a 52-minute spread. Employed Americans gained 32 minutes of sleep over the period; non-employed Americans gained just 19. Work is still the great sleep compressor, but it compresses less than it used to.

Drill down further and the numbers sharpen. Full-time workers (35+ hours per week) slept 8.70 hours in 2024, up from 8.15 in 2003 — a 33-minute gain. Part-time workers slept 9.18 hours, up from 8.66. But the most revealing cut is full-time workers on days they actually worked. In 2003, a full-time worker clocked just 7 hours and 38 minutes of sleep on a work day. By 2024, that had risen to 8 hours and 11 minutes — a gain of 34 minutes. This isn’t just about more vacation days or more people working from home on Fridays. It’s about workers sleeping more even on the days they go to work.

Remote work is an obvious candidate explanation. Before 2020, roughly 5% of workdays were done from home. After the pandemic, that share jumped to 25–30% and stayed there. Eliminating a 30-minute commute each way hands a worker a full hour — and the data suggests a meaningful fraction of that hour goes straight into the pillow. But the gains pre-date remote work. From 2003 to 2019, before COVID changed everything, full-time workers on work days had already gained 18 minutes of sleep. Whatever is driving this, it started well before Zoom.

Americans have always slept more on weekends, but the gap is closing. In 2003, the average person slept 8.26 hours on a non-holiday weekday and 9.29 hours on weekends and holidays — a difference of 62 minutes. By 2024, weekday sleep had risen to 8.80 hours and weekend sleep to 9.59, shrinking the gap to 47 minutes. Weekdays gained 32 minutes; weekends gained only 18.

This is significant because the weekday–weekend sleep gap is often used as a proxy for “sleep debt” — the idea that people chronically under-sleep during the work week and try to make it up on Saturday and Sunday. Sleep researchers have argued for years that weekend catch-up sleep is both widespread and inadequate. The narrowing gap suggests the debt itself is shrinking. Americans aren’t just sleeping more overall; they’re distributing their sleep more evenly across the week.

Two forces are probably at work. First, the rise of remote and hybrid work blurs the weekday–weekend boundary. When you don’t need to commute on Tuesday, Tuesday morning starts to look more like Saturday morning. Second, the broader cultural shift toward valuing sleep — the Arianna Huffington effect, the explosion of sleep-tracking devices, the Peloton-and-recovery wellness complex — affects weekday behavior more than weekend behavior, because weekdays had more room to improve. Nobody was cutting their weekend sleep short.

If employment is the single biggest predictor of how much you sleep, parenthood is the second. Parents of children under 18 slept 8.63 hours in 2024, versus the national average of 9.04 — a penalty of about 25 minutes per day. But the penalty varies dramatically by the age of the youngest child.

Parents whose youngest child is under 6 slept 8.72 hours in 2024. Parents of children 6 to 12 slept 8.67 hours. And parents of teenagers 13 to 17 slept just 8.40 hours — the least of any parental group. This is counterintuitive. Infants are the ones who wake up crying at 3 a.m. Toddlers are the ones who need to be carried back to bed. Why do parents of teenagers sleep the worst?

The answer lies in the composition of the group. Parents whose youngest child is 13–17 are overwhelmingly in their late 40s and 50s — the age bracket that already sleeps the least. They are at the peak of their careers, often in demanding managerial or professional roles. Many also have older children in college, adding financial stress. And the teenagers themselves are a factor: a 16-year-old who borrows the car and isn’t home by midnight produces a different kind of sleeplessness than a colicky infant. The ATUS can measure time in bed but cannot measure worry.

What the trend data does show is that the parent penalty is shrinking. Parents of children under 18 slept 8.34 hours in 2003 and 8.63 in 2024, a gain of 17 minutes. That’s less than the national gain of 28 minutes, so parents are falling behind in relative terms. But the absolute improvement is real. New parents in 2024 get more sleep than new parents in 2003 — perhaps because fathers do more nighttime caregiving than they used to, or because parenting norms have shifted toward accepting help and sharing the load.

The ATUS records behavior, not motivation. It can tell us that Americans sleep 28 more minutes than they did two decades ago, but it can’t tell us why. Still, four explanations are consistent with the data.

The work-hours decline. As we’ll explore in Episode 3, work hours have fallen across the period. The average American spent 3.67 hours per day working in 2003 and 3.43 hours in 2024. That’s 14 fewer minutes of work per day. Where did those minutes go? Sleep captured a substantial share. When people gain free time, they don’t necessarily spend it all socializing or exercising — they spend it sleeping. The body takes what it can get.

Remote work and the vanishing commute. For the roughly 30% of workers who now work from home at least part of the time, the elimination of a commute frees up 30–60 minutes per day. Some of that goes to work itself, some to household tasks, but a large fraction appears to go to sleep. The sharp jump from 8.84 (2019) to 8.95 (2021) hours — coinciding with the mass shift to remote work — is the clearest fingerprint of this effect.

The cultural revaluation of sleep. The sleep-wellness movement of the 2010s may have actually worked. Books like Why We Sleep, the rise of Oura rings and Apple Watch sleep tracking, the proliferation of weighted blankets and melatonin gummies, employer-sponsored wellness programs, and even corporate nap pods at companies like Google and Nike all contributed to a cultural message: sleep is not laziness, it’s performance optimization. Whether these specific products changed behavior or simply reflected an existing trend is impossible to untangle. But the timing aligns.

The aging population. America is getting older. The median age rose from 35.3 in 2000 to 38.9 in 2023. Older Americans sleep more than younger ones (until about age 65, where it plateaus). As the Baby Boomers moved from their 50s to their 60s and 70s over this period, the population shift alone would have pushed the national average upward. But this explains at most a few minutes of the 28-minute gain. The age-specific data shows that within each age group, sleep increased substantially.