Why Loneliness Surveys Keep Finding the Young Are Lonelier Than the Old

Gallup’s 142-country survey found that adults aged 19 to 29 report loneliness at 27%, compared with 17% among those 65 and older. Harvard’s Making Caring Common project found 61% of Americans aged 18 to 25 reporting serious loneliness, against a national figure of 36%. Cigna’s 2025 report found Gen Z and Millennials lonelier than Gen X and Baby Boomers, despite the younger cohorts having grown up with, and continuing to use, far more communication technology. AARP’s 2018 survey of adults 45 and older found the opposite gradient within that older population — loneliness fell with age within the 45-plus band. Put these together and a pattern holds across instruments, countries, and years: younger adults report more loneliness than older ones, consistently, using measures that were not designed to agree with each other.

That consistency is worth taking seriously, because loneliness surveys usually do not agree. The UCLA Loneliness Scale, single-item self-report questions, and Gallup’s “lonely a lot of the day yesterday” measure ask different things and produce different prevalence numbers as a result. When several instruments diverge on absolute levels but converge on a direction — younger reports more than older — that convergence is doing real evidentiary work, even if the exact size of the gap should not be taken from any one survey in isolation.

The harder question is why. At least four explanations circulate in the literature, and they are not mutually exclusive.

The reporting-bias explanation

One possibility is that the gap is partly an artifact of who is willing to say “lonely.” Older cohorts may under-report loneliness because of generational stigma around admitting it, while younger respondents, more fluent in the vocabulary of mental health, may be more willing to name the feeling on a survey even at equivalent underlying levels. This is plausible and impossible to rule out with cross-sectional self-report data alone. None of the sources reviewed here directly test it, which is itself a gap worth naming: a good study would need to compare self-report against a behavioral or physiological correlate across age cohorts, not just ask the question twice.

The life-stage explanation

A second account holds that this is not a generational shift at all but a life-stage effect that would show up in any era’s twenty-somethings. Young adulthood involves leaving school-structured social contact, moving cities for work, and often living alone or with strangers for the first time, all before the more stable social infrastructure of marriage, children’s schools, or long-tenure employment has formed. On this account, today’s 22-year-olds are lonelier than today’s 70-year-olds for the same reason that 22-year-olds have probably always been lonelier than 70-year-olds, and the AARP finding that loneliness declines with age within the 45-plus population is consistent with a smooth lifelong curve rather than a cohort break. This explanation predicts that today’s young adults will look like today’s older adults once they age into stable networks. It is testable only with longitudinal cohort data that follows the same people over decades, which none of the cross-sectional surveys cited here provide.

The cohort-specific explanation

A third account says something did change for people who came of age after roughly 2010, and the life-stage story cannot fully explain it because the size of the youth-loneliness gap looks larger now than earlier data would predict. Jonathan Haidt’s argument, developed at length and with extensive supporting documentation compiled jointly with Jean Twenge, is that the shift from play-based to phone-based adolescence altered the developmental environment for a specific cohort in a way that a simple life-stage account does not capture. This argument is contested. Several researchers dispute the strength of the causal claim, and the compiled evidence review itself includes dissenting studies rather than presenting the case as settled. What can be said without taking a side in that dispute is narrower: if a cohort-specific mechanism is operating, it should show up as an unusually large or persistent age gap relative to what older cross-sectional data implied, not merely as young adults reporting more loneliness than old adults, which the life-stage account already predicts. The Gallup and Harvard figures are consistent with either story on their own.

The American Enterprise Institute’s 2021 survey adds a data point that fits the cohort account somewhat better than the life-stage account. It found that 15% of American men report no close friends, a fivefold increase since 1990, and that the share of men with six or more close friends fell from 55% to 27% over the same period. A pure life-stage explanation would not predict a decline in a fixed benchmark like “close friends” measured the same way across three decades; it would only predict that young adults have fewer close friends than middle-aged adults at any point in time. A generational decline in the stock of close friendships, if real, points toward something structural rather than something that resolves itself as a birth cohort ages.

The infrastructure explanation

A fourth account is less about individual psychology and more about the physical stock of places available for informal contact. Research on third-place availability across the United States found closures in every category tracked — coffee shops, libraries, recreation centres, restaurants — between 2019 and 2021, concentrated in socially vulnerable and rural census tracts. Coverage of that research frames the loss of senior centres and churches as a driver of loneliness rather than a downstream symptom of it. This account does not obviously explain an age gradient by itself, since third-place loss affects all ages in a given geography. But it interacts with age: younger adults who have not yet accumulated stable networks through marriage, homeownership, or long employment tenure are more dependent on informal, place-based contact to build a network in the first place, whereas older adults draw more on networks already established. If that is right, third-place scarcity would hit the young harder even without any change in youth psychology.

Separately, the American Time Use Survey evidence on declining face-to-face socializing shows the decline predates the pandemic and continued afterward, which weakens any account that rests solely on COVID-era disruption. Gallup’s 2024 workplace data adds a related wrinkle: loneliness is higher among employees under 35 and higher again among fully remote workers than on-site ones, suggesting that the structure of a young person’s daily contact, not just their stated feelings, differs measurably from an older worker’s.

What would settle it

None of these four explanations is disqualified by the current evidence, and they are not competitors so much as candidates that would need to be isolated separately. A study capable of doing that would need to follow the same individuals from adolescence into midlife across at least two distinct birth cohorts, measuring the same instrument at the same life stages, while also tracking each person’s local third-place availability and close-friend count over time. That would separate a life-stage effect, which should look identical across cohorts at matched ages, from a cohort effect, which should not. No survey reviewed here does this. What exists instead is a set of cross-sectional snapshots that agree on the direction of the age gap and disagree, unhelpfully, on almost everything about what produces it.