How many close friends does the average person have?

It is more common than most people assume. Pew Research Center data from 2023 found that approximately 9% of men and 8% of women in the US report having no close friends. This does not mean it is ideal or without consequence, extensive research links social isolation to adverse health outcomes including increased all-cause mortality, higher cardiovascular risk, and increased rates of depression. But the prevalence of no close friendships means it is far from unusual, and it is something that can be addressed with intentional effort, even in adulthood.

Based on Pew Research 2023, n=4,618 · Survey Center on American Life 2021 · Dunbar 1992

Multiple explanations are proposed. Men’s friendships are more likely to be activity-based (formed around shared activities like sport or work) than women’s friendships, which are more likely to be disclosure-based (formed around emotional sharing). This means men’s friendships are more dependent on continued shared context: when the shared activity ends (a job change, a sport stopping), the friendship often fades. Women’s disclosure-based friendships can be maintained across greater distances and longer gaps. Men are also less likely to reach out to friends proactively when experiencing difficulties, creating a negative spiral where friends drift and emotional support becomes unavailable when needed.

Yes, though it is harder than in adolescence and early adulthood. Research by Jeffrey Hall at the University of Kansas found that it takes approximately 50 hours of time together to move from acquaintance to casual friend, 90 hours to become a friend, and over 200 hours to become a close friend. Adult life offers fewer unstructured hours and more competing obligations, making accumulating these hours challenging. The most effective approaches involve repeated, low-stakes contact in a shared context (a class, a club, a regular activity) combined with gradually increasing self-disclosure. Intentionality, actively scheduling time rather than waiting for connection to happen, is the variable most consistently associated with successful adult friendship formation.

Robin Dunbar's original 1992 paper in the Journal of Human Evolution proposed that the size of the human neocortex relative to the rest of the brain predicts a cognitive limit of approximately 150 on the number of stable social relationships a person can maintain simultaneously. This figure, now widely known as Dunbar's number, does not mean 150 close friends; it means 150 people whose identity, relationship to you, and relationship to others in your network you can actively track. In Dunbar's updated 2021 work, the 150-person active network is the fourth layer of a five-layer model. The innermost support clique of approximately 5 people and the sympathy group of approximately 15 are the layers most relevant to wellbeing and crisis support. The 150 figure is better understood as the upper ceiling of a functional community than as a friendship target.

Loneliness is a subjective experience of unmet social connection, which means it is possible to be lonely with many acquaintances and not lonely with very few close relationships. The US Surgeon General's 2023 advisory on the loneliness epidemic cited data showing that approximately 50% of American adults report measurable loneliness, defined as regularly feeling isolated, left out, or lacking meaningful connection. This figure coexists with widely varying friendship counts: some people with large social networks feel profoundly lonely because the connections are shallow, while others with one or two close relationships report high social satisfaction. The strongest predictor of loneliness in research is not network size but relationship quality, specifically the perceived availability of emotional support when needed.

Quality consistently outperforms quantity in the research. A 2017 meta-analysis by Holt-Lunstad et al. (Perspectives on Psychological Science, 148 studies, over 300,000 participants) found that social integration, which captures both quantity and quality, predicted mortality, but when the components were separated, relationship quality was the stronger predictor of all-cause mortality and psychological wellbeing. The Harvard Study of Adult Development, one of the longest running longitudinal studies of adult life (80+ years of follow-up), reached the same conclusion: the quality of close relationships at midlife was a stronger predictor of health and happiness at 80 than cholesterol levels, exercise habits, or career success. Having even one high-quality close relationship appears sufficient to protect against the health effects of social isolation.

Both transitions are associated with measurable friendship attrition. Marriage typically involves a consolidation of social networks as the couple's joint social life partially replaces individual social lives, and time available for maintaining friendships independent of the partnership declines. The transition to parenthood has a more pronounced effect: the Survey Center on American Life (2021) found that parents were significantly more likely than non-parents to report declining friendship counts in the preceding five years. Time constraints are the primary mechanism, with the average new parent losing 400 to 700 hours annually to infant care that previously went to other activities including social contact. The friendships most likely to survive the parenthood transition are those with other parents, where shared experience and similar life schedules maintain regular contact, and those with deep enough roots that occasional contact sustains the relationship.

The research suggests they can be but typically are not equivalent in the specific functions most relevant to wellbeing. Fardouly et al. (2018) and subsequent studies found that online social contact satisfies some components of social need, particularly information-sharing, weak-tie maintenance, and ambient social presence, but is less effective than in-person contact for the emotional co-regulation, physical comfort, and deep disclosure that characterise support clique friendships. A 2021 meta-analysis by Mitev et al. found that video-call social interaction was more effective than text-based online interaction in producing wellbeing effects, but both fell short of equivalent in-person contact. Crucially, online friendship does not appear to substitute for in-person friendship in the support clique layer: people with large online networks but few in-person close friends show loneliness rates comparable to those with small networks overall.

The pandemic accelerated pre-existing trends in friendship decline while adding acute losses. A 2021 survey by the American Enterprise Institute found that 18% of Americans reported losing a close friendship during the pandemic period, citing either direct conflict (particularly over health-related disagreements) or natural lapsing from lack of in-person contact. The Survey Center on American Life (2021) documented the sharpest single-year drop in close friendship counts in the time series, with the percentage of Americans reporting no close friends rising from 3% in 1990 to 12% in 2021. Recovery has been partial: subsequent surveys suggest some relationship re-engagement but not a full return to pre-pandemic friendship density for many adults, with those who were geographically isolated during restrictions showing the most persistent losses.

The health consequences of social isolation are substantial and well-documented. Holt-Lunstad et al.'s 2015 meta-analysis (148 studies, 308,849 participants) found that adequate social relationships were associated with a 50% increased likelihood of survival compared to those with poor social relationships, controlling for age, sex, and health status at baseline. The effect size is comparable to quitting smoking and exceeds the effects of obesity, physical inactivity, and excessive alcohol use as a mortality risk factor. Mechanistically, loneliness and social isolation are associated with elevated cortisol, increased inflammatory markers including IL-6 and CRP, dysregulated sleep, and reduced immune function. The US Surgeon General's 2023 advisory, the first of its kind on loneliness, cited these figures in framing social isolation as a public health crisis comparable in impact to obesity and physical inactivity.

Cross-cultural variation in friendship expectations is substantial. In many East Asian cultures, particularly in China, Japan, and South Korea, friendship is typically understood as a smaller, more exclusive category than in North American or Northern European contexts, but within that category, obligations and intimacy expectations are higher. What counts as a friend versus an acquaintance varies: research by Adams and Plaut (2003, Personal Relationships) found that Ghanaian adults defined friendship more narrowly and assigned it more practical support functions than American adults. In Mediterranean and Latin American cultures, friendship networks tend to be larger and more family-integrated, with the boundary between friend and extended family being more permeable than in Northern European or American contexts. These differences mean that cross-cultural comparisons of friendship counts are partially artefacts of definition rather than genuine differences in the depth of social connection.