Why the Loneliest People Are Not the Oldest

Gallup’s 2023 survey of 142 countries found that 27% of adults aged 19 to 29 report feeling very or fairly lonely, against 17% of adults 65 and older — the lowest rate of any age group measured. This is not an anomaly. Harvard’s Making Caring Common project found 61% of Americans aged 18 to 25 reporting serious loneliness in 2021, compared with far lower rates among older cohorts. Cigna’s 2025 report on loneliness in America found Gen Z and Millennials reporting more loneliness than Gen X and Baby Boomers, despite the older groups having lived through more years without smartphones, social media, or, presumably, less technological connectedness. Across three separate survey instruments, in three different years, the same age pattern holds: loneliness runs downhill as people get older, not uphill.

This sits awkwardly next to two decades of policy infrastructure built substantially around the opposite assumption. The UK’s 2018 national loneliness strategy, the first of its kind by any government, drew heavily on concern about isolated older adults. The National Academies of Sciences, Engineering, and Medicine devoted its landmark 2020 consensus report specifically to older adults, estimating that roughly one quarter of adults 65 and older are socially isolated. AARP has surveyed adults 45 and older on loneliness since 2018 and again in 2025, tracking a rise in loneliness within that population — but by definition never comparing it to the age group the Gallup and Harvard data say is worse off.

None of this is wrong, exactly. It is measuring a different thing.

Isolation is not loneliness, and age moves them in opposite directions

The National Academies figure — a quarter of older adults socially isolated — describes a structural property of a person’s network: how many people they have contact with, how often, and across how many types of relationship. The Gallup and Harvard figures describe a subjective state: whether a person feels lonely, regardless of how large or small their network actually is. Holt-Lunstad’s 2015 meta-analysis in Perspectives on Psychological Science treated these as related but distinct risk factors, reporting separate mortality odds ratios for social isolation, loneliness, and living alone. They correlate. They are not interchangeable, and this is the single most common source of confusion in reporting on the field.

A 2024 study in Scientific Reports looked directly at how the relationship between isolation and loneliness varies by age during the pandemic period and found that the two constructs decouple as people get older. Older adults, on average, have objectively smaller and more constrained networks — fewer coworkers, fewer parents-of-young-children peer groups, more widowhood, more mobility limitation. By the isolation metric, they score worse. But they report feeling lonely less often. Younger adults show the reverse: large nominal networks, frequent digital contact, more people in the room — and yet the highest subjective loneliness of any cohort in nearly every survey that measures it.

Part of the explanation is likely adjustment of expectation. A person who has restructured their life around a smaller, more stable set of relationships may not experience that smallness as deprivation. A young adult navigating unstable friendships, new cities, and early career pressure may have more social contact by volume and still feel that none of it is dependable. The 2025 HINTS-6 analysis in the American Journal of Preventive Medicine, using 2022 data, found that over 37% of U.S. adults report moderate-to-severe loneliness — but that study, like most national surveys, does not report the isolation-loneliness gap by age cleanly enough to settle which mechanism dominates.

What this means for where the resources go

If the goal is preventing isolation-related health harm — the elevated mortality risk Holt-Lunstad’s meta-analyses document — older adults remain the right target, because isolation itself, independent of how it feels, carries risk. If the goal is reducing loneliness as a subjective experience, the survey evidence says the intervention priority should skew younger, not older. These are not the same policy problem, and treating loneliness surveys as evidence for isolation programs, or vice versa, misallocates attention regardless of which age group actually needs help.

The comparability problem compounds this. AARP’s 2018 and 2025 surveys use the same 20-item UCLA Loneliness Scale, which makes their older-adult trend genuinely trackable over time. Gallup’s global figures use a different single-item measure. Harvard’s 2021 survey used its own instrument. None of these differences invalidate the age pattern — it recurs too consistently across too many instruments to be measurement noise — but they mean no single number should be treated as the loneliness rate for any age group. What would settle the mechanism question is a longitudinal panel that measures both structural isolation and subjective loneliness on the same individuals as they age, rather than comparing different age groups at a single point in time. That study does not yet exist in the published literature.