Homophily is the principle that contact between similar people occurs at a higher rate than among dissimilar people (reviewed in McPherson et al., 2001). It is a long-established principle (see e.g., Byrne (1969), which has 1000+ citations), and is particularly well established for friendships, which have been shown to be homophilous for ethnicity, age, religion, education, and social values ([16,156–160] in Dunbar (2018)). ‘Baseline’ homophily occurs because of the demography of the potential relationship pool, and effects beyond that occur due to other factors not limited to personal preference.

Homophily acts on a variety of different dimensions. The ‘Big 5’ personality traits of Agree-ableness, Extraversion, and Openness are similar among friends (university freshmen in the Netherlands) (Selfhout et al., 2010). Physical characteristics of age, height, weight, body fat, and hand-grip strength are homophilic in camp-mate and gift networks among the Hadza hunter-gatherer group (Apicella et al., 2012), and humans generally trend to trust people with faces who look like our own (cited in Sutherland et al., 2020). There is evidence for neural homophily: friends are exceptionally similar in how they perceive, interpret, and react to the world around (Parkinson et al., 2018), and socially close individuals have similar patterns of endogenously-driven attentional states over time (Hyon et al., 2020). There is evidence for genetic homophily (and heterophily): of 6 genes tested that were known to have significant effects on human social behaviour, one associated with alcoholism was homophilic among friends, and one associated with Openness was heterophilic (Fowler et al., 2011). There is also some evidence that homophily acts on cooperative/altruistic traits (below).

Homophily also occurs in other animals. In other primates: male macaques with similar personalities (gregariousness) formed stronger social bonds (Ebenau et al., 2019); juvenile rhesus monkeys with similar activity level, kinship, and sex have more temporally stable friendships (Weinstein and Capitanio, 2012); and chimpanzees with similar personalities (Sociability, Boldness, and Grooming Equity) were more likely to be friends (sitting in contact with each other) (Massen and Koski, 2014). These results support the hypothesis of a shared evolutionary origin of personality homophily as partner choice strategy in human and non-human primates (Ebenau et al., 2019). Homophily in other animals is reviewed in Weinstein and Capitanio (2012). One example is that dolphins show homophily based on their foraging tactic (cooperatively with human fishers) beyond what would be expected from the shared activity (i.e., when travelling, socialising, and resting) (Machado et al., 2019).

Why be homophilic?

Aside: Why even have friends?

A lot of the homophily work focuses on friendships. One of the challenges of explaining friendship is that people often, at high cost to themselves, lend assistance to a friend when they are in the most desperate need. This is best illustrated by the Banker’s Paradox (Tooby et al., 1996): when you desperately need a loan then you are a bad credit risk and a bank is less likely to give you one. Similarly, if friendship is about reciprocal altruism, then we should be least likely to extend assistance when our friend is most in need. Yet the opposite is true, and indeed we are offended by the idea that our friend should pay us back.

One way to resolve the Banker’s Paradox is by friends becoming indispensable to each other (Cole and Bruno Teboul, 2004). When individuals are in highly interdependent relationships, they create extensive mutual knowledge together. Therefore, one may be willing to lend considerable assistance to a friend in dire need, not based on the narrow premise that a friend might ultimately pay back such a generous favour, but as a means of protecting a precious and hard to replace resource.

One notion of partnership — that an individual should be willing to pay a small survival cost to increase the survival of their friend — can be modelled using a neighbour-modulated inclusive fitness approach (Eshel and Shaked, 2001; Rodrigues and Kokko, 2016).

Psychological explanations

Some explanations for homophily are focused on individual psychology. For example, the reinforcement-affect model proposes that similarity reinforces an individuals’ already-existing opinions, views, and feelings and thereby triggers an implicit affective response that increases attraction (Clore & Byrne (1974) cited in Selfhout et al., 2010). Other explanations involving mutual understanding, ease of communication, and shared goals and values may be related to solving coordination problems (see next section).

However, psychological explanations for homophily don’t necessarily rule out their evolutionary origin in cooperation problems. Cole and Bruno Teboul (2004) argues human behaviour and reasoning are built on an emotional foundation, and that psychological functioning has been shaped by problems encountered by our ancestors, including cooperation problems. For example, feeling understood is an important indicator of intimacy and relational satisfaction, but mutual understanding also improves coordination. Sedikides and Skowronski (2000) argues that the symbolic self may have partly evolved so that one could know their role and the group’s expectations of the self, and that self-esteem — which depends on group members’ evaluations of the self – may facilitate behaviour that benefits the group. Therefore, the emotions may provide an adaptive proximate mechanism for cooperative behaviour, and the logic underlying the behaviour can be covert.

There is evidence that humans have evolved cognitive specialisations for reasoning about social exchange. One nice example is that conditional reasoning is easier if a problem is framed in terms of social exchange and rule breaking (Sugiyama et al., 2002). Further evidence for the domain-specific nature of social reasoning comes from experiments performed on patients with damage to particular parts of the brain. Stone et al. (2002) found that a patient with damage to orbitofrontal cortex and amygdala performed poorly in Wason selection tasks that were framed as social contract problems compared to tasks framed as hazard problems (this was significantly different to control subjects, who performed equally well on each).

Homophily facilitates coordination

One commonly cited reason why individuals should choose partners homophilically is because it facilitates coordination between them on shared tasks. Cole and Bruno Teboul (2004) proposes that the ideal partnership is heterophilic (complementary) with respect to resources, skills, and abilities, but homophilic with respect to goals, beliefs, values, and interests. In particular, people tend to go with what they know, so coordination is easier when knowledge and behaviour are similar; and similarity increases insight into the other’s view, so one can guess their intention and behaviour. Similarly, uncertainty reduction theory proposes that homophily provides predictability, allowing individuals to communicate with less effort and greater confidence, which also makes coordination easier (Berger & Calabrese (1975) cited in Selfhout et al., 2010).

McElreath et al. (2003) modelled the coordination argument for homophily explicitly as a coordination game. They argued that their model allowed the evolution of markers because, in contrast to the ‘green beard’ model when applied to social dilemmas, when the payoffs have a coordination structure, there is nothing to be gained from cheating.

Homophily as a generalisation of kin preference

Cooperation with non-kin might occur when kinship benefits are misapplied to proximate others. For example, when discussing the finding that we tend to trust people with faces that look more like our own, Sutherland et al. (2020) suggested that this perhaps occurred for kinship reasons. However, the idea that friendship is kinship misapplied seems unlikely for reasons explained in Cole and Bruno Teboul (2004), e.g., because we have a long history of making distinctions between kin and non-kin. Kinship facilitates bond formation in macaques, but it is not a prerequisite for it (De Moor et al., 2020).

Here, we might need to be careful about what we conceive of us ‘kin’ and the degree of relatedness that entails. In small societies of 100–200 people, almost everyone is ‘kin’ either by descent or by marriage, and individuals who do not meet these criteria may be given honorary kin status (Barnard (1978) cited in Dunbar, 2018).

By this expanded definition of ‘kin’, the markers of kin-group membership are very similar to the key homophily dimensions that predict friendship closeness (Dunbar, 2018): language/dialect, where you grew up, educational history, hobbies and interests, sense of humour, and worldview (moral, religious, political) (Curry and Dunbar, 2013). Noteworthy too is the fact that most of these markers are cultural (Dunbar (2018) disagrees with literature above that homophily on personality attributes is as important).

Dunbar (2018) argues that certain kinds of cultural traits are particularly useful as markers for a cooperative group. Dialect above is acquired early in life and difficult to mimic. It is also something that can change relatively rapidly over generations. Citing the modelling work of Nettle and Dunbar (1997), which found that rapid evolution in the tag was necessary for cooperators to exclude defectors (similar to Antal et al. (2009)), Dunbar (2018) argues that cultural markers like dialect can change quickly so that invading defectors are always one step behind the rest of the community.

Homophily on cooperative traits themselves and positive assortment

Similarity not only triggers attraction but also triggers altruistic and cooperative behaviour. For example, similarity between friends is a predictor of expressed willingness to perform an altruistic act (donate a kidney) (Curry and Dunbar, 2013). Perceived similarity generally triggers empathy (i.e., a disposition to imagine how one would feel in another’s place). In an early experiment to test if empathy mediates altruism, Krebs (1975) found that subjects who were told that they were similar in personality and values to a roulette player reacted more strongly when he received electric shocks for losing and were more willing to help the performer at a cost to themselves.

Earlier, we discussed homophily on personality and other dimensions, but there is some evidence that homophily occurs on cooperative traits themselves. In Hadza hunter-gatherer groups, individuals who were more willing to donate a stick of honey were more likely to receive sticks of honey and to be chosen as camp-mates (Apicella et al., 2012). Further, this effect extended to two degrees of separation, suggesting that there might be polarisation in a social network such that cooperators tend to be connected to other cooperators and non-cooperators to other non-cooperators.

One mechanism by which cooperative-trait homophily may arise is by preferential tie formation and dissolution. Rand et al. (2011) found evidence for this mechanism in an experimental game involving a dynamic network. They also found that cooperation was promoted provided that the participants in the game were able to update their social ties rapidly enough (only the opportunity for frequent updating was needed, but the network itself did not necessarily change frequently). Similarly, reciprocated friendships in juvenile rhesus monkeys are more likely to persist (Weinstein and Capitanio, 2012).

It is interesting to note that some of the earlier mentioned personality dimensions of homophily are also associated with cooperation. Selfhout et al. (2010) observed homophily on the Agreeableness personality trait, which is mainly expressed through altruistic behaviours (Denissen & Penke, 2008a). They also observed that Agreeable individuals are more desirable as friends; however, from a game-theoretic perspective, it only makes sense to be altruistic if one’s partner is altruistic too, which implies that Agreeable individuals should seek out other Agreeable partners. One possible explanation for homophily on the Extraversion trait is that it is about matching partners’ needs for proximity and social sensitivities (Ebenau et al., 2019); however, high Extraversion in chimps is also correlated with sensitivity to inequality (Brosnan et al., 2015), which is suggestive of a matching of expectations in social exchange. Interestingly, Brosnan et al. (2015) also found that low Agreeableness is correlated with sensitivity to inequality in chimps, which seems to be a trait that would be useful for a long-term reciprocal relationship.

If homophily can act directly on the cooperative trait, then that has implications for the evolvability of cooperation. If cooperators tend to form ties with other cooperators, then that leaves defectors with no choice but to form ties with other undesirable noncooperators. Some authors argue that it is positive assortment itself which is the fundamental mechanism by which helping behaviours can evolve (e.g., Fletcher and Doebeli, 2009).

Arbitrary traits, minimal groups, and parochial cooperation

The discussion above concerns homophilic traits that are somehow meaningful, such as traits that have some functional role in maintaining relationships; however, there is evidence that humans will respond positively to similarity in traits that carry little meaning at all. For example, Burger et al. (2004) found that undergraduates who believed that they had the same birthday or similar fingerprints as a requester were more likely to comply with the request. As another example, people will give higher attractiveness ratings to similar others regardless of whether the similarity is on an significant attribute (politics and religion) or a relatively unimportant one (attitudes to gardening) (citations in Billig and Tajfel, 1973).

Pioneering work to establish the minimal conditions for biased behaviour was performed by Henri Tajfel and colleagues. In their most famous experiments (Tajfel et al., 1971), they categorised boys into groups based on similarity in (1) whether they over- or underestimated the number of dots on a screen, and (2) their preference for one or another painting. They then asked the boys to allocate money to other participants, who were anonymous except for an ID number and which group they were in. The key finding was that the majority chose to allocate money in a way that favoured their own group (experiment 1), and furthermore, that the boys were more concerned with creating a difference that favoured their own group than maximising the amount their group received (experiment 2).

It should be noted that the context of this work was quite specific: Tajfel was a Jewish prisoner-of-war, and the work was primarily interested in group categorisation itself, rather than (perceived) similarity. In subsequent work, they removed the role of similarity altogether by making the categorisation explicitly random: based on the flip of a coin (Billig and Tajfel, 1973). They replicated the biasing effect, and further observed that random categorisation into groups had a stronger effect than when there was a division based on ‘real’ similarity (paintings preference) but the division was not explicitly interpreted in terms of ‘groups’ (i.e., the word ‘group’ was not used in the instruction sheet, and the ability of participants to identify similar others based on their ID numbers was described as being incidental).

Tajfel’s work led to the concept of social identity theory (Tajfel and Turner, 1979). In short, the theory is based on three assumptions:

  1. individuals strive to maintain or enhance self-esteem;
  2. self-esteem is influenced by social identity; and
  3. evaluation of the social identity group is done in reference to other groups.


  1. individuals strive to maintain or increase social identity;
  2. positive social identity is determined by favourable comparisons to relevant outgroups;
  3. if social identity is unsatisfactory, individuals will either strive to leave the group or make their existing group more positively distinct.

In the case of the minimal group experiments described above, the only differentiation possible was monetary gain, and that’s why boys allocated money in a way that would maximise the difference between them.

Similarity-based groups and mere categorisation influences behaviour in collective action problems. Kramer and Brewer (1984) found that both real and a randomised minimal group treatments increased restraint in a common-pool resource game, particularly as the resource became more imperilled. The finding was replicated in Brewer and Kramer (1986), and it was compared with behaviour in a public goods game (contributing instead of harvesting), which revealed the additional complication that group identity had the opposite effect (decreasing cooperation) when the PGG group was large. Interestingly, both of the experiments above used what’s called a ‘common fate’ treatment to induce the minimal group. The idea is that the ‘group’ is induced by whether the same coin-flip was used to determine the payment all participants would receive (points-to-dollars conversion) or if they were determined by separate coin flips. Even though the outcome was the same between groups (the coins always ‘randomly chose’ to pay participants 5c per point), this was enough to induce the difference in behaviour.

Tajfel’s minimal group paradigm was specifically designed to remove any instrumental reasons for group bias; however, meta-analysis of subsequent work experimental work suggests that behaviour in various games is better explained by bounded generalised reciprocity (Balliet et al., 2014). Bounded generalised reciprocity (BGR) is the idea that group situations set-up the expectation that other members of the group will reciprocate one’s cooperation (Yamagishi and Mifune, 2008). Within the group, the mechanism maintaining cooperation is indirect reciprocity, which is maintained by within-group reputation.

The general theme of experiments is to compare BGR and social identity explanations by manipulating the possibility or expectation of reciprocation. For example, Jin and Yamagishi (1997) compared favours given to an ingroup member when the recipient knew the donors group-membership status versus not (reciprocity not possible). In the social identity theory, the only goal is to differentiate groups, so the recipient’s knowledge should not matter; but they found that ingroup bias did not occur in the latter treatment. As another example, Mifune et al. (2010) compared how ingroup bias (dictator game) responded to exposure to stylised eyes (Haley and Fessler, 2005). As expected from the indirect reciprocity component of BGR, ingroup bias was evidence in the watching-eye treatment but not without.

BGR fits into a larger concept of parochial cooperation, which is that humans behave in an ingroup bounded fashion, displaying both ingroup love and outgroup hate, facilitated by oxytocin (De Dreu et al., 2014). In this theory, we have a ‘group psychology’ that causes us assign ourselves and others to groups, and identify with and form our self-concept from them. This psychology is a consequence of our evolutionary history: BGR as discussed above, but also our history of intergroup conflict (authors seem to invoke a vague, group-selection type argument?). As a consequence of the BGR component, humans should have a general depersonalised trust in the ingroup, expect to receive benefits from other ingroup members, and be motivated to establish and maintain a good reputation as cooperator in ingroup.


Antal, T., Ohtsuki, H., Wakeley, J., Taylor, P. D. and Nowak, M. A. (2009). Evolution of cooperation by phenotypic similarity, Proceedings of the National Academy of Sciences 106(21): 8597–8600.

Apicella, C. L., Marlowe, F. W., Fowler, J. H. and Christakis, N. A. (2012). Social networks and cooperation in hunter-gatherers, Nature 481(7382): 497–501.

Bahns, A. J., Crandall, C. S., Gillath, O. and Preacher, K. J. (2017). Similarity in relationships as niche construction: Choice, stability, and influence within dyads in a free choice environment., Journal of Personality and Social Psychology 112(2): 329.

Balliet, D., Wu, J. and De Dreu, C. K. (2014). Ingroup favoritism in cooperation: A meta-analysis., Psychological Bulletin 140(6): 1556.

Billig, M. and Tajfel, H. (1973). Social categorization and similarity in intergroup behaviour, European Journal of Social Psychology 3(1): 27–52.

Brewer, M. B. and Kramer, R. M. (1986). Choice behavior in social dilemmas: Effects of social identity, group size, and decision framing., Journal of personality and social psychology 50(3): 543.

Brosnan, S. F., Hopper, L. M., Richey, S., Freeman, H. D., Talbot, C. F., Gosling, S. D., Lambeth, S. P. and Schapiro, S. J. (2015). Personality influences responses to inequity and contrast in chimpanzees, Animal behaviour 101: 75–87.

Burger, J. M., Messian, N., Patel, S., Del Prado, A. and Anderson, C. (2004). What a coincidence! the effects of incidental similarity on compliance, Personality and Social Psychology Bulletin 30(1): 35–43.

Byrne, D. (1969). Attitudes and attraction, Advances in experimental social psychology 4(1969): 35–89.

Cole, T. and Bruno Teboul, J. (2004). Non-zero-sum collaboration, reciprocity, and the preference for similarity: Developing an adaptive model of close relational functioning, Personal Relationships 11(2): 135–160.

Curry, O. and Dunbar, R. I. (2013). Do birds of a feather flock together?, Human nature 24(3): 336–347.

De Dreu, C. K., Balliet, D. and Halevy, N. (2014). Parochial cooperation in humans: forms and functions of self-sacrifice in intergroup conflict, Adv. Motiv. Sci 1: 1–47.

De Moor, D., Roos, C., Ostner, J. and Schülke, O. (2020). Bonds of bros and brothers: Kinship and social bonding in postdispersal male macaques, Molecular Ecology 29(17): 3346–3360.

Dunbar, R. I. (2018). The anatomy of friendship, Trends in cognitive sciences 22(1): 32–51.

Ebenau, A., von Borell, C., Penke, L., Ostner, J. and Schülke, O. (2019). Personality homophily affects male social bonding in wild Assamese macaques, Macaca assamensis, Animal Behaviour 155: 21–35.

Eshel, I. and Shaked, A. (2001). Partnership, Journal of Theoretical Biology 208(4): 457–474.

Fletcher, J. A. and Doebeli, M. (2009). A simple and general explanation for the evolution of altruism, Proceedings of the Royal Society B: Biological Sciences 276(1654): 13–19.

Fowler, J. H., Settle, J. E. and Christakis, N. A. (2011). Correlated genotypes in friendship networks, Proceedings of the National Academy of Sciences 108(5): 1993–1997.

Haley, K. J. and Fessler, D. M. (2005). Nobody’s watching?: Subtle cues affect generosity in an anonymous economic game, Evolution and Human behavior 26(3): 245–256.

Hyon, R., Kleinbaum, A. M. and Parkinson, C. (2020). Social network proximity predicts similar trajectories of psychological states: evidence from multi-voxel spatiotemporal dynamics, NeuroImage 216: 116492.

Kramer, R. M. and Brewer, M. B. (1984). Effects of group identity on resource use in a simulated commons dilemma, Journal of personality and social psychology 46(5): 1044.

Krebs, D. (1975). Empathy and altruism, Journal of Personality and Social psychology 32(6): 1134.

Machado, A. M. d. S., Cantor, M., Costa, A. P., Righetti, B. P., Bezamat, C., Valle-Pereira, J. V., Simões-Lopes, P. C., Castilho, P. V. and Daura-Jorge, F. G. (2019). Homophily around specialized foraging underlies dolphin social preferences, Biology letters 15(4): 20180909.

Massen, J. J. and Koski, S. E. (2014). Chimps of a feather sit together: chimpanzee friendships are based on homophily in personality, Evolution and Human Behavior 35(1): 1–8.

McElreath, R., Boyd, R. and Richerson, P. (2003). Shared norms and the evolution of ethnic markers, Current anthropology 44(1): 122–130.

McPherson, M., Smith-Lovin, L. and Cook, J. M. (2001). Birds of a feather: Homophily in social networks, Annual Review of Sociology 27(1): 415–444.

Mifune, N., Hashimoto, H. and Yamagishi, T. (2010). Altruism toward in-group members as a reputation mechanism, Evolution and Human Behavior 31(2): 109–117.

Nettle, D. and Dunbar, R. I. (1997). Social markers and the evolution of reciprocal exchange, Current Anthropology 38(1): 93–99.

Parkinson, C., Kleinbaum, A. M. and Wheatley, T. (2018). Similar neural responses predict friendship, Nature communications 9(1): 1–14.

Rand, D. G., Arbesman, S. and Christakis, N. A. (2011). Dynamic social networks promote cooperation in experiments with humans, Proceedings of the National Academy of Sciences 108(48): 19193–19198.

Rodrigues, A. M. and Kokko, H. (2016). Models of social evolution: can we do better to predict who helps whom to achieve what?, Philosophical Transactions of the Royal Society B: Biological Sciences 371(1687): 20150088.

Sedikides, C. and Skowronski, J. J. (2000). On the evolutionary functions of the symbolic self: The emergence of self-evaluation motives, in T. A., R. B. Felson and S. J. M. (eds), Psychological perspectives on self and identity, American Psychological Association, pp. 91–117.

Selfhout, M., Burk, W., Branje, S., Denissen, J., Van Aken, M. and Meeus, W. (2010). Emerging late adolescent friendship networks and big five personality traits: A social network approach, Journal of personality 78(2): 509–538.

Stone, V. E., Cosmides, L., Tooby, J., Kroll, N. and Knight, R. T. (2002). Selective impairment of reasoning about social exchange in a patient with bilateral limbic system damage, Proceedings of the National Academy of Sciences 99(17): 11531–11536.

Sugiyama, L. S., Tooby, J. and Cosmides, L. (2002). Cross-cultural evidence of cognitive adaptations for social exchange among the shiwiar of ecuadorian amazonia, Proceedings of the National Academy of Sciences 99(17): 11537–11542.

Sutherland, C. A., Burton, N. S., Wilmer, J. B., Blokland, G. A., Germine, L., Palermo, R., Collova, J. R. and Rhodes, G. (2020). Individual differences in trust evaluations are shaped mostly by environments, not genes, Proceedings of the National Academy of Sciences 117(19): 10218–10224.

Tajfel, H., Billig, M. G., Bundy, R. P. and Flament, C. (1971). Social categorization and intergroup behaviour, European journal of social psychology 1(2): 149–178.

Tajfel, H. and Turner, J. C. (1979). An integrative theory of intergroup conflict, in W. G. Austin and S. Worchel (eds), The Social Psychology of Intergroup Relations, Brooks-Cole, Monterey, CA, chapter 3, pp. 33–47.

Tooby, J., Cosmides, L. et al. (1996). Friendship and the Banker’s Paradox: Other pathways to the evolution of adaptations for altruism, Proceedings-British Academy, Vol. 88, OXFORD UNIVERSITY PRESS INC., pp. 119–144.

Weinstein, T. A. and Capitanio, J. P. (2012). Longitudinal stability of friendships in rhesus monkeys (Macaca mulatta): individual-and relationship-level effects., Journal of Comparative Psychology 126(1): 97.

Yamagishi, T. and Mifune, N. (2008). Does shared group membership promote altruism? fear, greed, and reputation, Rationality and Society 20(1): 5–30.