Cohort, Period, or Life Stage: Why Loneliness Skews Younger

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. Harvard’s Making Caring Common survey, fielded in the United States in late 2020, put youth loneliness even higher: 61% of adults 18 to 25 reported serious loneliness, compared with much lower rates among older respondents. The two figures are not directly comparable — different instruments, different years, different populations — but the direction is the same across nearly every dataset available: loneliness is highest among the young and lowest among the old. That is the opposite of the popular image of an isolated elderly person, and it deserves more scrutiny than it usually gets.

The pattern is robust enough that it is worth asking not whether it exists, but why. At least four distinct explanations circulate in the literature, and they carry very different implications for what, if anything, should be done about it.

Explanation one: life stage, not generation

The simplest account holds that young adulthood is structurally lonelier because it involves repeated network disruption. People in their late teens and twenties are leaving school, moving cities, changing jobs, and forming or dissolving romantic partnerships at higher rates than any other age group. Robin Dunbar’s account of social network structure — an inner layer of roughly five intimates, then 15, then 50, then 150 more distant contacts — implies that rebuilding those layers after a move or a graduation takes years, not months. On this account, elevated youth loneliness is a transitional cost that resolves as networks stabilize, and it would look similar in any generation experiencing similar transitions at the same age.

This explanation predicts that today’s 45-year-olds were also lonelier at 22 than they are now. Cross-sectional data like Gallup’s cannot test that prediction, because it compares different people at different ages in the same year rather than following the same people over time. A life-stage account and a genuine generational shift produce an identical single-year snapshot; only repeated measurement of the same cohort as it ages can tell them apart.

Explanation two: a genuine cohort effect

A competing account holds that something about growing up in the past two decades — smartphone-mediated socializing, delayed marriage and homeownership, the erosion of in-person institutions — has left younger cohorts with objectively thinner networks than their predecessors had at the same age. The American Enterprise Institute’s 2021 survey of American friendship offers some support: the share of Americans reporting zero close friends rose from 3% in 1990 to 12% in 2021, and the share of men with six or more close friends fell from 55% to 27% over the same period. That is a within-cohort trend over three decades, not a single age comparison, and it is consistent with a real generational shift in network size rather than a transitional dip.

If this account is right, current 25-year-olds will not simply “grow out of” elevated loneliness the way a 25-year-old in 1995 might have. Distinguishing this from the life-stage story requires a study that tracks specific birth cohorts across multiple decades, holding age constant and letting calendar year vary — a design almost none of the available surveys attempt, because most are single cross-sections repeated with different samples.

Explanation three: instrument and reporting differences by age

A third possibility is not that younger people have thinner networks, but that they report loneliness more readily. Cacioppo’s framing of loneliness as an aversive signal akin to hunger — evolved to prompt reconnection — says nothing about how comfortable different generations are naming that signal to a survey interviewer. Older cohorts, socialized in eras with more stigma around admitting emotional distress, may underreport. Younger cohorts, raised amid more open discussion of mental health, may be more willing to check the box.

This is difficult to rule out because most of the surveys driving the pattern use different instruments for different populations. The AARP Foundation’s 2018 survey of adults 45 and older used the 20-item UCLA Loneliness Scale, a validated academic instrument, and found that one in three older Americans qualify as lonely by that measure. Harvard’s youth-focused survey used its own items rather than the UCLA scale. When two age groups are measured with two different tools, a gap between them may partly reflect what each tool is sensitive to, not just what each population feels. The methods note here is a familiar one: an instrument mismatch between studies does not resolve a discrepancy, it explains why the discrepancy cannot be resolved with the data at hand.

Explanation four: survivorship in the mortality literature itself

There is a fourth, narrower wrinkle specific to the health literature rather than the prevalence surveys. Holt-Lunstad’s 2015 meta-analysis in Perspectives on Psychological Science found that the effect of social isolation and loneliness on early mortality was stronger in samples with an average age under 65 than in older samples. That is a finding about who dies from disconnection, not about who reports feeling lonely, and the two should not be collapsed into one story. One candidate explanation is selective survival: older adults who remain in a sample by a given wave are disproportionately those whose social deficits did not kill them earlier, which can flatten an association in older cohorts for reasons that have nothing to do with resilience or coping. The National Academies’ 2020 consensus report, which estimates that roughly one quarter of adults 65 and older are socially isolated by structural measures, is itself silent on why isolation’s mortality effect appears to weaken with age — a gap the report does not close.

What would settle it

None of these four explanations is disqualified by the current evidence, and they are not mutually exclusive. A defensible synthesis might combine all four in different proportions. What is missing is the one study design that could apportion the effect between them: a panel that follows identified individuals across at least two decades, using a single consistent instrument, across at least two distinct birth cohorts observed at the same ages. The Japanese JACSIS panel, which tracked isolation and loneliness longitudinally from 2020 to 2021, points toward the right design but covers too short a window and one country to answer a question about generational change. Until a study of that shape exists, the honest description of “why young adults report more loneliness than the old” is not an answer but a shortlist, and the size of each item on it is still unknown.