Homophily n.
Two mechanisms drive it: choice homophily, an explicit preference for similar others, and induced homophily, the uneven opportunity to meet anyone different in the first place.
/həˈmɒf.ɪ.li/
The definition
Homophily is the tendency of people to form social ties with others who share similar traits, such as age, education, ethnicity or belief. The principle structures relationships across friendship, marriage, work and advice networks alike, and much of it is set before any individual chooses a friend, through the institutions and settings that put similar people in each other's path.
The mechanism
The homophily principle, that similarity breeds connection, structures social ties across relationship types including friendship, marriage, work, and advice networks.1 Two distinct mechanisms produce the pattern. Choice homophily is a direct preference for people like oneself, visible when someone selects friends who share a background, an outlook, or an interest. Induced homophily works differently: it comes from the uneven distribution of opportunities to meet, set by shared social foci such as schools, workplaces, neighbourhoods, and clubs that put some people in each other's path far more often than others.2
Formal economic modelling shows that homophily can emerge from search costs and meeting opportunities alone, with no explicit preference for similarity required at all.3 Under that model the pattern is structural rather than psychological: people form ties with whoever sits nearest within their existing networks, and those networks are already sorted by circumstance before any preference operates. One consequence follows directly from the mathematics. Minority groups end up forming disproportionately fewer cross-group ties than majority groups do, not through any preference for segregation but because the pool of available majority contacts is larger relative to the pool of available minority ones.3
In practice
Consider how a professional network takes shape inside a large organisation, without anyone deliberately screening for similarity.
Worked example
A manager restructures a team and lets people self-select who they work alongside for a new project. Within weeks, the group that forms mirrors the manager's own background: similar tenure, similar training, similar communication style. No one intended to exclude anyone. The self-selection process simply favoured whoever felt easiest to work with, and ease correlated with similarity.
The result is a team that is easier to manage day to day but narrower in perspective than the organisation intended.
Why it matters
Homophily and social influence produce similar-looking clustering, and studies that fail to separate the two have overestimated peer influence in behaviour adoption by 300 to 700%.4 Anyone reading a claim about a contagious habit, a viral trend, or a peer-driven behaviour change should ask whether the underlying study controlled for the fact that similar people already cluster together before any influence has a chance to occur.
The same clustering plays out at scale online. Analysis of more than 100 million social media posts on controversial topics found that homophilic clustering dominates user interaction patterns on Facebook and Twitter, reinforcing echo chamber formation.5 That clustering is not uniform: it is pronounced on Facebook and Twitter but comparatively weaker on Reddit and Gab, where interaction structures differ, which means platform design itself shapes how far homophily narrows what a user sees.5
Questions of record
What is homophily in simple terms?
Homophily is the tendency for people to form social ties, such as friendships, marriages or working relationships, with others who resemble them in traits like age, education or belief. It operates across almost every kind of relationship and is one of the most consistently observed patterns in network science.
How is homophily different from social influence?
Homophily means similar people end up connected in the first place; social influence means connected people become more similar over time. The two produce almost identical clustering in observed data, which is why studies that do not separate them have overestimated peer influence in behaviour adoption by as much as 300 to 700%.
Does homophily cause echo chambers on social media?
It is a major contributor. Large-scale analysis of social media posts on controversial topics found that homophilic clustering, users interacting mainly with similar others, dominates interaction patterns on platforms such as Facebook and Twitter, which reinforces echo chamber formation rather than exposing users to a wider range of views.
Why do minority groups end up more homophilous in their networks?
Economic modelling shows this can happen without any preference for segregation. When meeting opportunities are shaped by search costs alone, minority groups form disproportionately fewer cross-group ties simply because the pool of available majority contacts is larger relative to the pool of available minority ones.