What the Loneliness Averages Hide About Young Men

Fifteen percent of American men report having no close friends at all, a fivefold increase since 1990. That figure, from the American Enterprise Institute’s 2021 Survey Center on American Life, does not appear in most loneliness coverage, which tends to report a single national average and move on. It should appear more often, because it describes a subgroup whose trajectory looks nothing like the aggregate.

The headline numbers on loneliness are usually reported this way: roughly a quarter of adults worldwide feel lonely, according to Gallup’s 142-country survey from 2023, and about a third of U.S. adults 45 and older report being lonely, per the AARP Foundation’s 2018 national survey using the UCLA Loneliness Scale. Cigna’s 2020 index put the U.S. adult figure at 61%, sometimes or always lonely, though that survey used a different instrument and a different threshold, which is a familiar problem in this literature and one worth flagging every time two numbers that look comparable turn out not to be. None of these breakdowns, on their own, tell a reader what has happened to men’s friendships specifically. That requires going to a source built for the question.

The AEI friendship data

The Survey Center on American Life’s 2021 report on American friendship is not, strictly, a loneliness survey. It asks about close friends: how many a respondent has, how often they see them, whether they talk about personal matters with them. That is a structural measure, closer to isolation than to loneliness in the distinction this Center returns to often — one describes the shape of a person’s network, the other describes how that person feels about it. The friendship data is nonetheless the clearest available picture of what has eroded and for whom.

In 1990, 3% of Americans reported having no close friends. By 2021, that figure was 12% for the population overall — a fourfold rise. Split by sex, the picture sharpens. Men went from 3% with no close friends in 1990 to 15% in 2021. Women’s rate rose too, but less steeply. The decline in large friend groups tells the same story from a different angle: 55% of men had at least six close friends in 1990; by 2021 that had fallen to 27%. Women’s equivalent numbers declined as well, but the drop among men is larger in absolute terms and starts from a higher base, meaning men lost more of what they had.

This is not a story that fits comfortably into either of the two loneliness narratives that dominate public discussion: the youth-loneliness story, built on data like Harvard’s Making Caring Common survey finding 61% of adults 18–25 seriously lonely, and the older-adult isolation story, built on the AARP and National Academies work on adults over 65. Middle-aged and older men are the group whose close-friendship networks contracted most, and they are largely absent from the loneliness surveys that get quoted in policy discussion, which skew toward youth and old age as the two poles of concern.

Why the average obscures this

A population-wide loneliness average is a weighted blend of very different trajectories. If men’s friendship networks have thinned dramatically while women’s have thinned modestly, and if young adults report acute loneliness for reasons distinct from either, then a single national percentage compresses at least three separate patterns into one number that describes none of them well. Gallup’s 2023 global data illustrates the same masking effect at the demographic level it does track: loneliness is highest among adults 19–29 at 27% and lowest among adults 65 and older at 17%, a generational gradient that a topline “24% of the world is lonely” erases entirely. The AEI data suggests a second gradient, by sex, running underneath the age one, and most surveys are not built to show it because they ask about feeling lonely rather than about the number and closeness of friends — the isolation measure that would actually reveal the male-specific decline.

Robert Putnam’s account of declining civic participation in Bowling Alone, published in 2000, anticipated a version of this pattern two decades before the AEI survey documented it in individual friendship terms. Putnam’s concern was associational life broadly — clubs, congregations, bowling leagues — and men were disproportionately represented in many of the institutions he tracked as shrinking, from union halls to fraternal organizations. The AEI numbers can be read as a later, more granular confirmation of a decline Putnam described at the institutional level: as the third places and associations that used to generate men’s friendships thinned out, the friendships built inside them appear to have thinned along with them.

Workplace data adds a partial corroboration

Gallup’s 2024 State of the Global Workplace found loneliness elevated among employees under 35 and among fully remote workers, at 25% compared with 16% for fully on-site employees. This is a different population and a different measure than the AEI friendship survey, and it does not break results down by sex in the published figures available here, so it cannot be used to confirm that remote work specifically drove the male friendship decline. It is worth citing mainly as a caution: if work is one of the venues where adult friendships have historically formed, and a growing share of adults now do that work remotely, the mechanism by which the AEI decline occurred becomes more plausible without becoming proven. The workplace figures describe a contemporaneous pattern, not a confirmed cause of the friendship trend, and the two surveys should not be merged into a single narrative.

What this does not establish

The AEI survey is cross-sectional at two time points, 1990 and 2021, not a continuous panel. It cannot show when between those years the decline occurred, whether it was gradual or concentrated around a particular period, or whether the pandemic accelerated a trend already under way — a distinction that matters a great deal for policy timing and that the data as reported cannot resolve. It also relies on self-report of friend counts, a measure vulnerable to how respondents define “close friend,” and that definition may itself have shifted over three decades in ways the survey cannot detect.

Nor does the survey measure loneliness directly, so it is not possible to say from this source alone whether men with fewer close friends report feeling lonelier as a result. The isolation-loneliness distinction cuts both ways here: a smaller network does not guarantee a lonelier subjective state, and some of the literature on this distinction, including the 2024 Scientific Reports study on isolation and age published during the pandemic, has found that the relationship between the two varies by age group in ways not yet worked out for sex.

What would settle the question

A longitudinal panel tracking the same male respondents’ friendship networks and loneliness scores across multiple points since 1990 would show whether the friendship decline is continuous, accelerating, or concentrated in one era, and whether it is actually driving subjective loneliness rather than merely coexisting with it. A survey using a consistent isolation instrument, cross-tabulated simultaneously by sex, age, and employment status, would also help separate the workplace-remote effect from the broader civic-associational decline Putnam described, rather than leaving the two as parallel but unconnected data points, as they currently are. Until such data exists, the specific claim that can be made is a narrower and more useful one than “men are lonely”: American men’s close-friendship networks have contracted more than women’s since 1990, on the AEI’s numbers, and that decline is invisible in every topline loneliness figure currently in circulation.