What Young Adults Say Versus What They Do

A 2021 Harvard survey found that 61% of adults aged 18 to 25 reported serious loneliness. A Gallup survey conducted globally in 2022 and 2023 put the figure for adults aged 19 to 29 at 27%. Both are described as measuring loneliness among young adults. The gap between them is not a contradiction to resolve. It is a reminder that “loneliness” is not one measurement, and neither instrument is describing the same thing as a count of how often young adults actually see other people in person.

That third figure — observed or administrative behavior, rather than a subjective rating — is the one this piece is interested in, because it is the one most often missing from the conversation.

Two different questions, asked two different ways

The Harvard Graduate School of Education’s Making Caring Common project surveyed Americans in the fall of 2020, capturing a moment of acute pandemic disruption to social contact. Its 61% figure for young adults reflects a period when in-person options were unusually constrained, and roughly half of lonely young adults in that survey reported that no one had asked how they were doing, in a way that felt genuine, in the recent weeks. That is a snapshot of a specific and unusual moment, not a stable trait of the age group.

Gallup’s estimate comes from a different design entirely: a global survey run across 142 countries between June 2022 and February 2023, well after acute pandemic restrictions had lifted almost everywhere. It found 24% of people worldwide report feeling very or fairly lonely, with young adults aged 19-29 highest at 27% and adults 65 and older lowest at 17%. The instrument, the timing, and the geography all differ from the Harvard survey. Neither number is wrong. They are not measurements of the same variable at the same time, and stacking them into a single “loneliness is up” or “loneliness is down” narrative flattens what each was actually built to capture.

The 2025 Cigna Loneliness in America survey adds a third data point, with fieldwork from late May to mid-June 2024: 57% of American adults overall report loneliness, and Gen Z and Millennial respondents report more loneliness than Gen X or Baby Boomers, despite reporting greater technological connectedness. That last clause is the interesting one. Self-reported loneliness and self-reported connectivity are moving in the same direction, not opposite ones, in the same population. If loneliness were simply the inverse of technological contact, this would not happen.

What the American Time Use data shows instead

Reporting from WBUR on American Time Use Survey data offers something closer to an observed measure: not how people rate their feelings, but how they actually allocate hours. That data shows face-to-face socializing has been declining in the United States for years, a decline that began well before COVID-19 and continued after pandemic disruptions eased. This is a behavioral record, collected through diary-style time accounting rather than a mood scale, and it points to something self-report loneliness surveys cannot directly show: a structural drop in the frequency of in-person contact, independent of how any individual feels about it.

The American Enterprise Institute’s 2021 survey of American friendship offers a parallel structural measure. It found that 12% of Americans report having no close friends at all, up from just 3% in 1990, and that the share of men with six or more close friends fell from 55% in 1990 to 27% by 2021. These are counts of network size, not ratings of subjective feeling, and they are the kind of measure that can be tracked and compared across three decades in a way that a single-year mood survey cannot.

This is the isolation-versus-loneliness distinction that the literature returns to repeatedly. Isolation is a structural property of a person’s network — how many people they see, how often, how many close ties they can name. Loneliness is the subjective gap between the connection someone has and the connection they want. A 2024 study in Scientific Reports examining the interplay between isolation, age, and loneliness during the pandemic found that the relationship between the two varies by age group, which is a strong argument against treating either as a proxy for the other. Someone can have a shrinking network and not feel lonely about it. Someone can have a nominally full calendar and feel intensely lonely anyway.

Why this matters for how the problem is framed

The debate over adolescent phones and mental health, laid out by Jonathan Haidt and elaborated in the collaborative evidence review he maintains with Jean Twenge, is instructive here because it is explicitly contested. Haidt argues that the shift to phone-based childhood in the early 2010s drove a measurable rise in loneliness and mental illness among adolescents. Other researchers dispute how much of that rise the phone shift can explain, and the evidence review itself includes dissenting studies rather than presenting a settled consensus. Self-report loneliness scales cannot adjudicate that dispute on their own, because they cannot distinguish a genuine drop in social contact from a genuine rise in dissatisfaction with contact that hasn’t actually changed. Only a behavioral measure — time-use data, network-size counts, observed contact frequency — can speak to the first half of that question.

A 2025 analysis of the CDC’s HINTS-6 survey data, collected in 2022, found that over 37% of U.S. adults experience moderate-to-severe loneliness, with roughly 1 in 7 in the severe category. That figure, like the others, describes a feeling. What would strengthen this literature is not another prevalence survey asking young adults to rate their loneliness, since that pool is now well populated. It is longitudinal behavioral data — time-use diaries, network rosters, administrative records of participation in clubs, teams, or congregations — tracked against the same individuals’ self-reported loneliness over the same years. Without that pairing, it remains possible to show that young adults feel worse and equally possible to show, separately, that they socialize less, without ever establishing which is driving which.