A much-repeated finding about intelligence and friendship sounds as if it gives clever people permission to withdraw. The underlying paper says something narrower. In a large American sample, frequent contact with friends was associated with slightly greater life satisfaction for most young adults, while the slope reversed at the upper end of a verbal-intelligence measure.
This is one study, not settled consensus. The data were correlational, collected in 2001–02, and the interaction was very small. It does not establish that intelligent people dislike company, need fewer relationships or would become happier by declining invitations.
The study used an existing American cohort
Norman Li of Singapore Management University and Satoshi Kanazawa of the London School of Economics published the analysis in the British Journal of Psychology in 2016. They used Wave III of the National Longitudinal Study of Adolescent to Adult Health, known as Add Health, rather than recruiting a new group for this particular paper.
The wave contained 15,197 people aged 18 to 28. The complete statistical model for friendship contact used 14,513 cases because some records lacked one or more variables. Add Health began with a nationally representative sample of American secondary-school students, but by Wave III these participants were young adults. The study’s published cohort profile explains the broader design.
Life satisfaction came from one question: “How satisfied are you with your life as a whole?” Responses ran from one to five. Socialising was also compressed into one item, asking how many times in the past seven days the respondent had hung out with friends or spoken to them on the phone for more than five minutes. The response scale ran from zero to seven or more.
Intelligence was measured with an abbreviated Peabody Picture Vocabulary Test. The raw results were converted to an IQ-style scale with a mean of 100 and a standard deviation of 15. The test captures verbal ability and correlates with general intelligence, but it is not a complete portrait of reasoning, creativity, knowledge or social skill.
The reversal was statistically real and practically tiny
When friendship contact was entered alone, it had no significant association with life satisfaction. Current marital status was masking the relationship: married respondents reported greater satisfaction on average while socialising with friends less frequently. Once marriage was included, more frequent socialising was positively associated with life satisfaction. The association remained after adjustment for age, sex, ethnicity and education.
The memorable result appeared in the interaction. As verbal-intelligence scores rose, the positive association between socialising and satisfaction weakened. The interaction coefficient was negative and statistically significant, with a p-value of 0.014.
For the authors’ illustrative higher-scoring group, one standard deviation above the mean at 115.57, predicted life satisfaction was 4.106 among frequent socialisers and 4.131 among less frequent socialisers. The difference ran against the overall pattern, but it was only 0.0248 points on a five-point scale. The authors estimated a standardised effect size of minus 0.03.
At the lower-scoring comparison point of 81.39, frequent socialisers were at 4.159 and less frequent socialisers at 4.116. That gap was 0.0423 points, with an estimated effect size of 0.05.
A large sample can identify a relationship this small with statistical confidence. Statistical significance and practical importance are different questions.
Higher scores did not produce a retreat from friendship
The behaviour in the dataset complicates the simple story. Respondents with higher verbal-intelligence scores actually socialised more frequently, with a correlation of 0.121. Whatever produced the slight reversal in the satisfaction model, it did not produce an obvious preference for less contact.
The paper did not measure introversion either. Introversion is a personality trait; performance on a vocabulary test is a cognitive score. Treating them as interchangeable turns two different constructs into a flattering stereotype about solitude.
I previously examined this dataset while unpacking that introvert-intelligence claim. The more defensible reading remains modest: one average relationship varied by verbal score in one young American cohort.
Frequency leaves out most of what makes friendship matter
One person can speak to a trusted friend twice in a week and feel understood. Another can have seven social encounters that are rushed, obligatory or emotionally thin. The Add Health item counts contacts, but does not tell us who initiated them, whether they happened one-to-one or in groups, how long they lasted beyond five minutes, or whether the respondent enjoyed them.
That missing texture matters because the wider literature does not reduce friendship to frequency. A 2023 systematic review of 38 adult-friendship studies found that friendship quality, support, maintenance, number of friends and socialising all appeared in the evidence on wellbeing. The overall pattern was positive, while the reviewers also identified sparse, contradictory and population-specific findings. Twenty-eight of the 38 studies focused mainly on young adults, often university students.
A separate pooled analysis of recorded daily conversations combined four samples totalling 486 adults. More conversation and more substantive conversation were moderately associated with life satisfaction. Personality did not substantially change those associations, and an earlier claim that small talk itself was harmful failed to replicate.
Neither source tells us that everyone needs the same social calendar. They do show why counting encounters alone is an unusually thin way to describe a relationship.
The proposed evolutionary explanation remains speculative
Li and Kanazawa framed their findings through what they called the “savanna theory of happiness”. Their proposal was that frequent contact with allies mattered in ancestral environments, while people with greater intelligence may be better able to cope with conditions unlike those in which human psychology evolved.
That theory is the authors’ interpretation, not something the Add Health variables directly measured. The dataset cannot observe ancestral adaptation, and it cannot establish why the modern interaction appeared.
Causal direction is unresolved as well. More contact might have slightly lowered satisfaction among some higher-scoring respondents. Less satisfied people might also have sought friends more often, or work, study and family circumstances might have required more contact. Unmeasured personality, health, employment and relationship quality could influence both sides of the association.
The finding argues against a universal social quota
The study’s most useful feature is its resistance to a single average prescription. A broadly positive association between friendship contact and life satisfaction can coexist with variation between people.
It cannot tell any individual how often to socialise. The analysis compared a one-week contact count with a global judgement of life recorded at the same wave, and the subgroup differences were only a few hundredths of a point. Stronger evidence would need to follow people over time, measure the quality and context of their relationships, and reproduce the intelligence interaction in other populations.
For now, the result is best treated as a small clue that social frequency may not mean the same thing for everyone, not as evidence that a particular kind of person thrives without friends.