Why Two Surveys Disagree on Youth Loneliness by Twenty Points

Two figures circulate in coverage of youth loneliness, and they are not close. Harvard’s Making Caring Common project reported in February 2021 that 61% of Americans aged 18 to 25 experienced serious loneliness. Gallup’s global survey, published in October 2023 and drawing on data collected between June 2022 and February 2023, put loneliness among 19- to 29-year-olds worldwide at 27%. Both are described in media coverage as “the” rate of loneliness among young people. They differ by more than a factor of two. Neither is wrong. They are measuring different things and calling the result by the same name.

What each study actually asked

The Harvard survey, run through Making Caring Common, sampled U.S. adults during the pandemic and asked about loneliness using questions designed to capture what the report calls “serious loneliness” — a category built from multiple markers of distress, not a single yes-or-no item. The 61% figure sits alongside a related finding from the same report: roughly half of lonely young adults said no one had spent more than a few minutes recently checking in on them in a way that felt genuine. The instrument was constructed for a report explicitly framed around the pandemic’s effect on connection, at a moment — February 2021 — when much of the country was still under significant social restriction.

The Gallup figure comes from a different instrument entirely, part of a 142-country survey conducted jointly with Meta. Respondents were asked a direct, single-item question about whether they felt lonely, with response options including “very,” “fairly,” “not much,” or “not at all.” Gallup’s 24% global loneliness figure, and the 27% figure specifically for adults aged 19 to 29, count only those who answered “very” or “fairly.” The companion Gallup–Meta report on social connectedness found 72% of respondents worldwide felt very or fairly socially connected — a majority, even among a population that also reports elevated loneliness by age. The two findings are not contradictory within that study; they describe different constructs sitting side by side in the same population.

Three sources of the gap

The disagreement between 61% and 27% has at least three separate causes, and it is worth separating them because they call for different kinds of caution.

First, the instrument. Harvard’s measure was built to detect a broader condition — “serious loneliness” — using multiple criteria, which tends to produce a higher headline number than a single self-rating question. Gallup’s single-item measure, by contrast, only counts people who chose the two most severe of four response options. A person who feels lonely “sometimes” or “a little” registers as not lonely in Gallup’s count but might register as lonely under a multi-item instrument. This is not a flaw in either design. It is a difference in what counts as the threshold for the category, and thresholds move headline numbers more than most other design choices.

Second, the population and moment. Harvard’s sample is U.S.-only, and the data were collected in the first year of the pandemic, when social contact for young adults — many of whom were in college or early career at the time — was unusually restricted. Gallup’s sample is global, collected roughly two years later, after most pandemic-era restrictions had lifted. A study documenting loneliness during acute isolation and a study documenting loneliness during a return to ordinary social contact are not measuring the same period of young adult life, even if both use the label “young adults.”

Third, the reference class. Harvard’s 18-to-25 band and Gallup’s 19-to-29 band overlap but are not identical age brackets, and neither report breaks its data down finely enough to isolate the effect of that difference on its own. It is a minor factor compared to instrument and timing, but it adds noise on top of the larger discrepancies.

Why this keeps happening

The pattern is not unique to these two studies. A 2023 review in BMC Public Health surveying the loneliness research literature identified inconsistent measurement as one of the field’s central obstacles to comparison — different studies use different instruments, different cutoffs, and different reference periods, and the resulting numbers get treated as interchangeable once they escape into public discussion. The AARP Foundation’s 2018 survey of adults 45 and older illustrates the alternative: it used the 20-item UCLA Loneliness Scale, a validated academic instrument, and found one in three older U.S. adults lonely. Because AARP used an established scale rather than a bespoke set of questions, its figure is more directly comparable to other academic literature using the same scale — even though it still cannot be compared cleanly to either the Harvard or Gallup figures, which use different instruments again.

This matters for anyone using these numbers in argument. A policy brief citing “61% of young adults are lonely” is making a claim about a pandemic-era U.S. population using a multi-item threshold. A brief citing “27% of young adults are lonely” is making a claim about a global population two years later using a single self-rating item. Both claims can be true and non-contradictory. Citing either one as a general, timeless fact about young adult loneliness — without naming the instrument or the year the data were collected — overstates what either survey actually supports.

What would resolve it

The cleanest fix is not a bigger survey. It is a survey that runs the same instrument, the same age brackets, and the same country coverage at two points in time, so any change reflects the population rather than the measuring stick. Absent that, the second-best option is what a handful of researchers now do explicitly: report loneliness prevalence alongside the specific instrument and cutoff used, so a reader can judge whether one figure is comparable to another before treating them as evidence of a trend. Right now, most public discussion of youth loneliness compares headline percentages without doing that work, which is how a single population can appear to be simultaneously in crisis and predominantly connected, depending only on which report is on screen.

Until studies converge on a shared instrument — or at minimum report their thresholds plainly enough for readers to adjust — the honest description of youth loneliness prevalence is a range bounded by measurement choice, not a single number. That range currently runs from roughly one in four to roughly three in five, and the size of that range is itself the more informative finding.