In May 2023, U.S. Surgeon General Vivek Murthy declared loneliness a public health epidemic, comparing its mortality risk to smoking 15 cigarettes a day. The declaration felt abstract — how do you measure loneliness? But the Bureau of Labor Statistics had been quietly tracking the answer for two decades. Since 2003, the average American has lost 12 minutes of daily socializing — a quarter of all face-to-face connection, gone. Fewer than 3 in 10 Americans now socialize on any given day, down from 4 in 10. And the decline was already steep long before anyone heard the word “lockdown.”
In 2003, the average American spent 47 minutes per day socializing and communicating — talking face-to-face with friends, calling family, hosting neighbors, catching up with coworkers in the break room. By 2024, that number had fallen to 35 minutes. Twelve minutes may sound trivial in the abstract. Multiplied across 330 million people and 365 days, it represents roughly 24 billion fewer hours of human connection per year.
The decline didn’t arrive suddenly. It unfolded like a slow leak. Socializing slipped from 0.78 hours per day in 2003 to 0.64 by 2019 — a steady 18% erosion over sixteen years, well before anyone locked a door for COVID. The pandemic then hammered the number down to 0.57 hours in 2021, and the recovery has been grudging: 0.56 in 2022, 0.57 in 2023, 0.59 in 2024. We are still 8% below pre-pandemic levels and 24% below where we started.
What makes this trend so striking is its relentlessness. There is no plateau, no year of reversal. The line moves in one direction, interrupted only by statistical noise and a brief uptick in 2012. Every other cultural shift — the iPhone launch in 2007, Facebook going mainstream in 2009, remote work after 2020 — corresponds with a further step down. The ATUS doesn’t explain causation. But it timestamps the decline with unusual precision.
The participation rate tells an even starker story. In 2003, 40.2% of Americans engaged in some form of socializing on any given day. By 2019, that had already dropped to 34.8%. After the pandemic, it collapsed to 29.1% and has barely recovered to 29.9% in 2024. Fewer than three in ten Americans now have a social interaction beyond the bare minimum required by work or errands on a typical day. That threshold — crossing below 30% — hasn’t been reclaimed since 2021.
The social calendar has always had a rhythm: weekdays are for work, weekends are for people. In 2003 that distinction was sharp — Americans socialized 1.18 hours on weekends versus just 0.61 hours on weekdays, a gap of 34 minutes. But the decline hit both halves of the week, and it hit weekdays harder.
Weekday socializing fell 28%, from 0.61 hours in 2003 to 0.44 in 2024. Weekend socializing fell 20%, from 1.18 to 0.94 hours. The gap has actually widened slightly — Americans are concentrating whatever social energy they have left into Saturday and Sunday. This mirrors a pattern we saw in Episode 3 with work: the American week is becoming more polarized, with weekdays squeezed ever tighter and weekends bearing more of the load for both rest and connection.
The post-pandemic numbers are telling. Weekday socializing crashed from 0.49 in 2019 to 0.43 in 2021 and has lingered at 0.42–0.44 ever since. Weekend socializing fell from 0.97 to 0.89 and has recovered to 0.94. The weekend is healing faster than the workweek. Remote work may have eliminated the office hallway conversation — the incidental socializing that doesn’t feel like socializing until it’s gone.
The loneliness epidemic is not evenly distributed. The youngest Americans — ages 15 to 24 — lost the most social time in both absolute and relative terms. In 2003, young adults socialized 1.01 hours per day, by far the most of any age group. By 2024, that had fallen to 0.62 hours — a loss of 23 minutes, or 39%. A generation that grew up with more communication tools than any in history spends less time actually communicating than any generation the survey has measured.
The decline among young people started well before the pandemic. By 2015, socializing for the 15–24 cohort had already dropped to 0.74 hours. By 2019, it was 0.65 — nearly as low as today’s level. COVID barely moved the needle for young adults because the needle had already moved. Sociologists like Jean Twenge have linked this to the smartphone era — her research marks 2012 as the year teen mental health and social behavior began diverging sharply, aligning almost exactly with when smartphone ownership passed 50% among U.S. teens.
Retirees (65+) experienced the second-largest drop in absolute terms: from 0.78 to 0.56 hours, a loss of 13 minutes. But the dynamics are different. Older Americans’ social time held steady through 2012 (0.77 hours) and even showed resilience through 2019 (0.67 hours). The pandemic then cut it to 0.51 hours in 2021, and the recovery to 0.56 has been sluggish. For seniors, COVID didn’t accelerate an existing trend — it created a rupture. Many of the institutions that structured retiree social life — church groups, community centers, volunteer organizations — either closed permanently or haven’t fully reopened.
Middle-aged Americans (35–54) showed the smallest declines. The 35–44 cohort went from 0.68 to 0.57 hours (−16%), while the 45–54 group dropped from 0.71 to 0.62 (−13%). These are peak career and parenting years, when socializing was already compressed. There was less to lose.
In 2003, women socialized more than men: 0.82 hours versus 0.73 hours, a gap of about 5 minutes. Women have historically been the social connectors — the ones who organize dinners, maintain friendships across households, call relatives, coordinate children’s playdates. By 2024, women’s socializing had fallen to 0.61 hours (−26%) while men’s dropped to 0.56 hours (−23%). The gap shrank from 5 minutes to 3.
The convergence isn’t because men started socializing more — men declined steadily throughout the period. It’s because women’s social time eroded faster. Between 2003 and 2024, women lost 13 minutes while men lost 10. This mirrors the pattern we found in Episode 4: women are increasingly carrying the same work burden as men while still doing more household labor, leaving less margin for the social ties they once maintained. When something had to give, connection was what gave.
The decline in socializing hit every demographic group, but the depth varied dramatically. Black Americans experienced the steepest fall: from 0.81 hours in 2003 to just 0.48 in 2024 — a 41% decline. In absolute terms, that’s a loss of 20 minutes per day. No other racial group lost even half as much. In 2003, Black Americans socialized the most of any group measured. By 2024, they socialized the least.
Hispanic Americans were the most resilient, dropping from 0.78 to 0.69 hours (−12%). Asian Americans fell sharply from 0.77 to 0.45 by 2019, but then rebounded to 0.57 in 2024 — one of the few demographic groups to show a meaningful post-pandemic recovery. White Americans declined steadily from 0.77 to 0.60 (−22%), tracking close to the national average.
Education tells a different story — one where less schooling correlates with more social connection. Americans without a high school diploma socialized 0.82 hours in 2003 and still managed 0.74 in 2024 — a decline of just 10%. High school graduates dropped from 0.76 to 0.55 (−28%), and college graduates fell from 0.69 to 0.57 (−17%). The less-educated have consistently socialized more, and the gap has widened. One interpretation: blue-collar work still happens in person, while knowledge work migrated to Zoom calls that end when the meeting ends.
| Group | 2003 | 2010 | 2019 | 2024 | Change |
|---|---|---|---|---|---|
| By Race/Ethnicity | |||||
| White | 0.77 | 0.69 | 0.66 | 0.60 | −22% |
| Black or African American | 0.81 | 0.74 | 0.58 | 0.48 | −41% |
| Hispanic or Latino | 0.78 | 0.77 | 0.70 | 0.69 | −12% |
| Asian | 0.77 | 0.61 | 0.45 | 0.57 | −26% |
| By Education (25+ yrs) | |||||
| Less than high school | 0.82 | 0.72 | 0.71 | 0.74 | −10% |
| High school graduates | 0.76 | 0.66 | 0.57 | 0.55 | −28% |
| Some college / associate | 0.69 | 0.64 | 0.72 | 0.59 | −14% |
| Bachelor’s degree+ | 0.69 | 0.67 | 0.61 | 0.57 | −17% |
| By Employment | |||||
| Employed | 0.70 | 0.65 | 0.61 | 0.53 | −24% |
| Not employed | 0.93 | 0.80 | 0.69 | 0.68 | −27% |
Employment status confirms an intuitive result: people who don’t work have more time to socialize. In 2003, the not-employed socialized 0.93 hours compared to 0.70 for the employed — a 33% premium. But even the not-employed have seen a 27% decline, suggesting that the forces driving social withdrawal are broader than just time pressure. Retired Americans, stay-at-home parents, the unemployed — everyone is socializing less, regardless of how much free time they have.
Socializing isn’t the only form of community engagement that eroded. Attending or hosting social events — dinner parties, barbecues, birthday gatherings — dropped from 0.10 hours per day in 2003 to just 0.05 in 2024, a 50% decline. That’s 3 minutes per day that sounds negligible until you realize it represents the structured occasions that turn acquaintances into friends: the dinner invitation extended, the backyard gathering planned, the holiday party hosted.
Volunteering held steadier, declining from 0.14 hours in 2003 to 0.13 in 2024 (−7%), though it cratered to 0.09 during the pandemic and has only partially recovered. Religious and spiritual activities were flat overall — 0.14 hours in both 2003 and 2024 — but dipped to 0.11 in 2021 and 2023, suggesting church attendance hasn’t fully returned to pre-pandemic rhythms even as the headline number appears unchanged.
One countertrend stands out: telephone calls rose from 0.08 hours per day in 2003 to 0.13 in 2024, with a spike to 0.16 in 2021 when it was the only option for millions. Phone calls are up 63% — but that gain of 3 minutes doesn’t begin to offset the 12 minutes lost from in-person socializing. Americans are substituting, but at a fraction of the rate needed to fill the gap.
Perhaps the most alarming metric in this dataset isn’t how long people socialize but whether they socialize at all. The ATUS tracks participation rates — the percentage of Americans who engage in an activity on any given day. For socializing and communicating, that rate has dropped from 40.2% in 2003 to 29.9% in 2024.
Think about what that means. On a typical day in 2024, more than 7 out of 10 Americans have zero social interaction beyond what’s required by their job or their errands. No phone call to a friend. No coffee with a neighbor. No conversation that isn’t transactional. The daily social fabric has thinned to the point where most Americans go through an entire day without a single voluntary social exchange.
The pre-pandemic trajectory was already concerning: 40.2% (2003) → 37.0% (2010) → 34.8% (2019). The pandemic then broke it below 30% for the first time — 29.1% in 2021 — and we have not climbed back above that threshold. The difference between 40% and 30% represents roughly 33 million fewer Americans socializing on any given day compared to 2003. That’s a population the size of Texas that simply stopped showing up.
| Measure | 2003 | 2010 | 2019 | 2024 | Change |
|---|---|---|---|---|---|
| All persons (hrs/day) | 0.78 | 0.70 | 0.64 | 0.59 | −24% |
| Participation rate (%) | 40.2 | 37.0 | 34.8 | 29.9 | −10.3 pp |
| By Day Type | |||||
| Weekdays | 0.61 | 0.55 | 0.49 | 0.44 | −28% |
| Weekends & holidays | 1.18 | 1.07 | 0.97 | 0.94 | −20% |
| By Gender | |||||
| Men | 0.73 | 0.68 | 0.62 | 0.56 | −23% |
| Women | 0.82 | 0.73 | 0.66 | 0.61 | −26% |
| By Age Group | |||||
| 15–24 years | 1.01 | 0.89 | 0.65 | 0.62 | −39% |
| 25–34 years | 0.77 | 0.79 | 0.65 | 0.56 | −27% |
| 35–44 years | 0.68 | 0.61 | 0.59 | 0.57 | −16% |
| 45–54 years | 0.71 | 0.67 | 0.59 | 0.62 | −13% |
| 55–64 years | 0.71 | 0.59 | 0.67 | 0.59 | −17% |
| 65+ years | 0.78 | 0.66 | 0.67 | 0.56 | −28% |
| Related Activities | |||||
| Social events | 0.10 | 0.08 | 0.06 | 0.05 | −50% |
| Telephone calls | 0.08 | 0.10 | 0.11 | 0.13 | +63% |
| Volunteering | 0.14 | 0.15 | 0.13 | 0.13 | −7% |
| Religious activities | 0.14 | 0.16 | 0.14 | 0.14 | — |
The ATUS offers what the Surgeon General’s report couldn’t: a precise measurement of social withdrawal, tracked year by year for two decades. Americans lost 12 minutes of daily socializing — a quarter of all face-to-face connection — and the participation rate fell below 30% for the first time in recorded history. The decline was broad-based, hitting every age group, every race, every education level, and both genders. But it was sharpest among young adults, Black Americans, and high school graduates.
The pandemic accelerated what smartphones started. But the forces behind this retreat are structural: remote work eliminated incidental contact, algorithmic feeds replaced phone calls, and a housing market that scattered people into longer commutes left less time for the neighbors next door. Twelve minutes a day is the price. Whether Americans can buy it back — or whether the new social equilibrium is simply lonelier — is the open question the data cannot yet answer.