What the data actually shows
The clearest U.S. picture comes from the Survey Center on American Life's American Perspectives Survey. Its 2021 data found the median adult reported around three to four close friends, with roughly half reporting three or fewer. The same survey tracked a notable shift over three decades: the share of Americans saying they had no close friends rose from about 3% in 1990 to around 12% in 2021, while the share reporting ten or more close friends fell sharply over the same period.
These figures fit a broader framework from anthropologist Robin Dunbar. His 'social brain' work suggests humans can sustain only a limited number of stable relationships — popularly summarised as 'Dunbar's number,' around 150 — arranged in layers. Within that, the data points to an inner circle of roughly five intimate relationships and about fifteen close ones. On that view, a small number of genuinely close friends is not a shortfall; it is close to the ceiling of what most people can maintain at once.
Time-use data adds context to the trend. The American Time Use Survey shows that average time spent socialising in person with friends has fallen over the past two decades. So the picture is less that people have failed at friendship and more that the conditions for it — unstructured shared time, in particular — have thinned, which makes maintaining even a small circle harder than it used to be.
Why this feels different from how it actually is
The most visible friendships are the loud ones. Group photos, big celebrations, and dense social feeds make other people's circles look large and effortless, while the quieter reality — a couple of close ties and a wider ring of acquaintances — is exactly what does not get posted. You are comparing the size of your real circle to a curated impression of everyone else's.
Friendships also thin in predictable ways after major transitions, and those transitions cluster in adulthood. Moving, changing jobs, becoming a parent, and the simple loss of the daily proximity that school and early work provide all tend to shrink the active circle. Because these changes feel personal, the resulting smaller circle can feel like a personal failing rather than a near-universal pattern tied to life stage.
There is also a definitional trap. 'Close friend' means something specific — someone you would confide in or call in a crisis — but in the moment people often measure themselves against the much larger count of acquaintances, followers, and contacts. Counted honestly, almost everyone's set of genuinely close friends is small, which is why the real median lands at three or four rather than the dozens the word 'friends' can suggest.
What the research says to do about it
The research on friendship leans consistently toward consistency: regular, low-stakes contact tends to matter more than grand gestures. Studies of how friendships form and survive point to repeated, unstructured time together as the main ingredient, which is why deliberately scheduling recurring contact — a standing call, a weekly walk — does more to sustain a friendship than waiting for the right occasion.
Depth appears to matter more than breadth. Because the inner circle is naturally small, the data suggests that investing in a few close ties generally does more for wellbeing than trying to expand the total count. Prioritising the handful of relationships that already feel close is well aligned with how human social networks are actually structured.
Initiative is usually underrated. Surveys repeatedly find that people underestimate how much others want to hear from them and how much a reached-out-to friend appreciates the contact. Acting on the impulse to reconnect — rather than assuming you would be intruding — is one of the few low-cost moves the evidence broadly supports.
What the research says does not help
Treating a high friend or follower count as the goal does not help, and can mislead. The data on close ties shows that the meaningful number is small by nature, so optimising for breadth tends to spread limited time too thin to sustain the few relationships that actually carry the weight.
Waiting for friendship to happen passively — assuming proximity or shared circumstances will do the work — tends to fail in adulthood, precisely because the structures that once supplied automatic contact have fallen away. The time-use trend suggests that without deliberate effort, even existing friendships quietly lapse.
Comparing your circle to other people's most visible social moments is actively distorting. Those moments are the highlight reel of connection; measured against them, almost any real circle will look thin, which produces a feeling of deficiency that the actual median — three to four close friends — does not support.
Real numbers in context
The honest reference point is modest. The median adult reports roughly three to four close friends, and about half report three or fewer (Survey Center on American Life, 2021). The share of adults reporting no close friends rose from around 3% in 1990 to about 12% in 2021 — a real and worth-noting shift, but one that still leaves the large majority with at least one close friend.
Dunbar's framework puts these numbers in perspective: an inner circle of about five intimate ties and roughly fifteen close ones is near the upper bound of what most people can sustain. Against that ceiling, three or four close friends is not a sign of having too few — it is squarely within the normal range, and the feeling that everyone else has more is largely an artefact of which friendships are visible.