Loneliness Skews Younger, Not Older

The National Academies’ 2020 consensus report on loneliness opens with a figure that has shaped a decade of policy attention: roughly one quarter of adults aged 65 and older are considered socially isolated. That statistic, and the broader body of work behind it, has anchored a widely shared assumption that loneliness is primarily a problem of old age — of widowhood, retirement, and shrinking social circles late in life.

The survey data on age and loneliness do not support that assumption nearly as cleanly as the policy focus implies.

What the surveys actually show

The AARP Foundation’s 2018 national survey of 3,020 adults aged 45 and older, using the 20-item UCLA Loneliness Scale, found that one in three respondents qualified as lonely. That is a substantial figure, and it is the number most often cited when discussing loneliness among midlife and older adults.

Compare it with the Harvard Graduate School of Education’s Making Caring Common project, which surveyed Americans in 2020 and found that 36% reported serious loneliness overall — but 61% of young adults aged 18 to 25 did, along with 51% of mothers with young children. Cigna’s 2020 workplace loneliness survey, fielded before the Harvard report, found 73% of workers aged 18 to 22 reporting loneliness, and more than 80% of employed Gen Z respondents overall.

Put those side by side and the pattern is consistent across three independently fielded surveys, using three different instruments, run by three different organizations: young adults report loneliness at roughly double the rate of the midlife-and-older population that has received the bulk of clinical and policy attention.

This is not a new observation buried in a single outlier study. It shows up whenever a survey asks a broad age range the same question in the same wave.

The instrument problem, stated plainly

Before drawing too firm a conclusion, the comparison needs a caveat that the Center returns to often: these figures are not directly comparable in the way a reader might want them to be.

The AARP survey used the UCLA Loneliness Scale, a validated 20-item instrument with a long research history, which is one reason its 33% figure sits comfortably within the range reported by academic studies using the same tool. The Harvard Making Caring Common survey and the Cigna survey used different question wordings and different response scales. A single-item “how often do you feel lonely” question and a 20-item multidimensional scale do not measure identical things, even when both are labeled loneliness, and they can produce different prevalence estimates for reasons that have nothing to do with the underlying reality.

That caveat matters, but it does not resolve the age pattern. The AARP survey itself, despite using the more conservative UCLA instrument, still found lower loneliness in its 45-and-older sample than either of the other two surveys found in their younger samples. If anything, the instrument difference should make the comparison more forgiving toward the older population, since the UCLA scale is generally considered a stricter and more clinically grounded measure than a single self-report item. The gap persists regardless.

The American Enterprise Institute’s 2021 survey on American friendship adds a related, less obviously “loneliness” data point: 12% of Americans report having no close friends at all, up from 3% in 1990, and the collapse is sharpest among men, whose share with six or more close friends fell from 55% in 1990 to 27% in 2021. That decline is generational as much as it is age-specific — it reflects a difference between cohorts, not simply what happens to a person as they age.

Why the mortality evidence complicates the picture further

Julianne Holt-Lunstad’s 2015 meta-analytic review in Perspectives on Psychological Science, covering social isolation, loneliness, and living alone as mortality risk factors, contains a finding that cuts directly against the assumption that these risks matter most for the old. The review reported that social deficits were more predictive of death in samples with an average age under 65 than in older samples. Isolation, loneliness, and living alone were each independently associated with increased mortality risk — odds ratios of 1.29, 1.26, and 1.32 respectively — but the strength of that association was not concentrated in the elderly population the public health conversation tends to picture.

This does not mean isolation is harmless for older adults. The National Academies’ report and the American Heart Association’s 2022 scientific statement both describe real and serious cardiovascular consequences of isolation and loneliness, including roughly a 30% increased risk of heart attack, stroke, or death from either, with worse outcomes for people who already have coronary heart disease or a prior stroke. Age remains a relevant risk factor for the downstream health consequences of isolation, in part because older adults carry more baseline cardiovascular and physiological vulnerability.

What the Holt-Lunstad review suggests is narrower and more specific: the relative predictive power of social deficits for mortality does not scale up with age the way a naive model of “loneliness gets worse and more dangerous as people get older” would predict. A younger person who is chronically isolated may be at comparatively higher relative risk than an older person with the same isolation score, even though the older person’s absolute risk of death from any cause is higher to begin with. Relative risk and absolute risk are being blurred together in a lot of public discussion of this topic, and the two point in somewhat different directions here.

Why the mismatch between attention and prevalence exists

Part of the explanation is structural rather than epidemiological. Older adults are easier to survey through existing channels — Medicare enrollment, senior centers, retirement communities — and they interact more regularly with a health care system that the National Academies’ report and its 2020 clinical commentary in the American Journal of Geriatric Psychiatry argue should be screening for isolation as a matter of routine practice. Young adults, by contrast, are dispersed across employers, universities, and no institution in particular, which makes them harder to reach with a formal intervention and easier to overlook in a policy document.

There is also a plausible mechanism difference. The AARP survey identified network size, network diversity, and physical isolation as the strongest predictors of loneliness among older respondents — 33% of those who had spoken with a neighbor recently were lonely, against 61% of those who never had. Isolation in later life is often structural: driving stops, a spouse dies, a social circle contracts through attrition. Loneliness in early adulthood looks different — it coexists with dense digital contact and often with cohabitation or new parenthood, as the Harvard survey’s finding on mothers of young children suggests. These may be distinguishable phenomena that happen to share a name and a measurement tradition, which is exactly the kind of conflation the loneliness literature struggles with generally: isolation as a structural fact about a person’s network, and loneliness as the subjective gap between the connection someone has and the connection they want. A young parent can be constantly surrounded by people and still report high loneliness on a survey instrument that is measuring something closer to unmet expectation than actual contact.

What would settle this

The clean answer would come from a single longitudinal cohort, tracked with one consistent instrument across the full adult age range over ten or twenty years, distinguishing structural isolation from subjective loneliness at every wave. No source in this comparison offers that. What exists instead is a set of independently fielded, differently instrumented cross-sectional surveys that happen to agree on direction even where they disagree on magnitude.

That agreement is worth taking seriously. It suggests that the near-exclusive association of loneliness with old age in public discussion is not wrong about older adults facing real and clinically significant risk — it is wrong about where the numerically larger problem currently sits.