Should birds of a feather flock together?

Background

You’ve probably heard the saying that “birds of a feather flock together,” and also the one about how “opposites attract.” The first of these statements speaks to our natural tendency to befriend those who are like ourselves and the second characterizes our pursuit of diversity. Given their widespread repetition, you might wonder which of these is correct. Moreover, the statements raise interesting questions about how we organize our social worlds, how we relate to those around us, and what impacts these decisions have on our individual and collective wellbeing. In multi-cultural contexts, the answers to these questions are deeply meaningful and have real impacts on individuals and communities. As such, understanding how we compose (or should compose) our social networks, with respect to diversity, is of critical import.

Purpose

The purpose of this evidence brief is to examine the literature related to social network composition, particularly with respect to the impact of socializing with others who are similar or different from ourselves. In doing so, we recognize that any one person may have a multitude of identities and that the intersection of these identities creates unique experiences (Crenshaw, 1998). As such, it is readily obvious that it is immensely difficult to gauge whether two individuals are similar or different (Kalantari, 2021). In practice, we rely on a limited subset of characteristics, which vary from arbitrary to significant (Ertug et al., 2022). While we recognize this limitation, the present brief nevertheless aims to characterize existing research related to social network diversity with hopes of being able to understand what is generally known and understood about this topic.

Evidence from Existing Studies

In considering social network diversity and the effect it has on our wellbeing, it is helpful to define its opposite. In the scientific literature, the term homophily describes the extent to which individuals in a social network are similar to one another across any given characteristic or set of characteristics, including socio-economic status, values, beliefs, or demographic traits (Currarini et al., 2009; Fu et al., 2012; McPherson et al., 2001). When individuals in a social network share a characteristic in common, the network is said to have high homophily. Conversely, when individuals in a network differ with respect to some characteristic, the network is said to have high diversity (Fazelpour & Rubin, 2022).

Describing Patterns of Homophily and Diversity

The existing literature has demonstrated that social networks do tend towards higher homophily across a wide range of demographic measures (e.g., ethnicity, age, religion, education, occupation, gender) and in-depth characteristics (e.g., sense of humour, moral beliefs, community belonging) of individuals (McPherson et al., 2001; Burleson et al., 2009; Curry & Dunbar, 2013). In noting this trend, we should be aware that it is only a trend and that in most contexts, individuals do differ from one another across a variety of characteristics. As such, real-world homophily should be understood only as a propensity for similarity across different characteristics. Furthermore, evidence suggests that there is a diminishing return on the number of shared features necessary to solidify social connections (Block & Grund, 2014) and that we share the greatest degree of homophily with the inner layers of our social network (Curry & Dunbar, 2013).

With respect to the causes of homophily, social selection likely plays an important role (Verbrugge, 1977; Hampton et al., 2018; Lieberman & Shaw, 2019; Sun & Taylor, 2020; Branas-Garza et al., 2022) – as is exemplified by the strong degree to which our most intimate social contacts are similar to ourselves. For example, Horwitz and colleagues (2023) show that across 133 trait correlations in sexual partnerships, 130 exhibited high degrees (i.e., >80%) of homophily. However, the strength of homophily is also contingent on the size of social groups and the broader population contexts in which social groups are situated. For example, larger groups tend to have greater diversity, while smaller groups may be more homogenous (Au, 2023; Currarini et al., 2009). Similarly, individuals who come from minority groups may have greater outgroup social ties than those coming from majority groups (Currarini et al., 2009).

Benefits of Homophily

In understanding the effects of homophily on individuals and communities, researchers have identified numerous benefits at the individual and group-levels. At the individual level, homophily fosters positive psychological states, including a sense of belonging and acceptance. For example, Van der Horst & Coffé (2011) demonstrated that interacting with similar others leads to a greater sense of belonging and acceptance, contributing to increased self-esteem and self-worth. This effect may result from a sense of shared understanding across like-minded individuals (Centola, 2011; Fu et al., 2012). In addition to the psychological benefits, Centola (2011) suggests that homophily across social networks can contribute to uptake of positive health behaviours and thereby contribute to improved physical health.

At the group level, homophily can also foster social cohesion, enhance within-group coping strategies and emotional support, and ease barriers to friendship formation (Au, 2023; Puga-Gonzalez et al., 2020; Reagans, 2010; Sias et al., 2008). For example, in exploring the effects on friendship formation Maarten et al (2009) reports that shared characteristics influence communication patterns and attraction between individuals, with more similarity making friendship formation easier. In these ways, homophily can be adaptive. For some minority communities (e.g., 2SLGBTQ+ community), this adaptation may be especially important – providing minority enclaves that serve as refuge from stigma, discrimination, and violence (Gillespie et al., 2015; Logan, 2013; Mollica et al., 2003). Taken together, these various benefits explain the observed propensity towards homophily.

Consequences of Homophily

However, in addition to the benefits outlined above, homophily can also negatively impact individuals and groups. It can lead to discrimination, social exclusion, and isolation – cutting individuals and their groups off from important outside perspectives, contributing to social inequalities, and reinforcing negative health behaviours (Au, 2023; Pampel et al., 2010; Schaefer et al., 2011; Centola, 2011; Rostila, 2010; Christakis & Fowler, 2007; Christakis, 2004; Friedman & Aral, 2001; McPherson et al., 2001). These negative effects can be particularly harmful when homophily is structurally imposed, such as in instances of racial segregation (Williams & Collins, 2001).

These various effects highlight the benefits of diversity. Indeed, intergroup contact can contribute to positive health outcomes. It can enhance mental health, physical well-being, and reduce mortality, particularly in older adults (Ali et al., 2018). Additionally, intergroup contact with people from diverse backgrounds can lead to a reduction in prejudice and stereotyping, as these interactions enhance empathy, cultural competencies, and attunement to different experiences and perspectives (Erickson, 2003; McFarland et al., 2014; McPherson et al., 2001; Sias et al., 2008). Further, building relationships with people from different demographics can lead to stronger social bonds and a more interconnected community. Such conditions create trust and cooperation between social groups, which is essential for affirmative action, fighting adversity, and promoting upward mobility for marginalized groups (Chetty et al, 2022).

Case Study: Can diversity damage social capital?

Several researchers have noted that having diverse ethnic neighborhoods might lead to lower social capital and trust, especially in the United States, while other countries have found mixed results. Given this predicament, Fieldhouse & Cutts (2010) aimed to compare the relationship between neighborhood ethnic diversity and social capital in the US and Great Britain across different ethnic groups. The researchers used individual and neighborhood census data and surveys to explore the relationship between people's backgrounds, their neighborhoods, and their sense of community. The results identified that for white folk, having a diverse neighborhood is connected to less trust in both countries. However, this association is not as strong for minority groups. The major takeaway here is that minority groups tend to respond differently to diversity in comparison to the White majority, that is, ethnic minorities tend to be more comfortable living in diverse areas even where that diversity is from people of other ethnic groups. Continued research is needed to understand these effects and how they can be mitigated.

Analyses from the Canadian Social Connection Survey

Using data from the 2022 Canadian Social Connection Survey, we asked a subsample of participants (n = 439) what proportion of their close friends were of similar age, gender, ethnicity, sexual orientation, geographic location, political orientation, religious orientation, and socioeconomic status. As shown in Figure 1, homophily was highest with simple demographic traits, particularly sexual orientation and gender. This trend likely reflects a combination of selection bias (e.g., choosing friends similar to oneself) and the natural population imbalances in Canada due to unequal group sizes (e.g., ~95% of people are heterosexual).

In addition to these analyses, we also examined whether having more close friends was associated with homophily across any of these characteristics. Results suggested that having more close friends was associated with greater homophily on religion (p = 0.006) and socioeconomic status (p = 0.0335) in bivariable analyses.

To understand the effect of homophily on mental health, we conducted regression analyses. The only dimension of homophily associated with differences in self-rated mental health was shared political beliefs (p = 0.0307), and the effect was negative, suggesting that homophily on this trait was associated with lower self-rated mental health. Based on these results, we sought to further explore the role of homophily using a case study approach. In doing so, we explored the effect of homophily on sexual orientation.

As a first step, we examined the prevalence of sexual orientation homophily across sexual orientation groups. This analyses showed that sexual orientation homophily was highest among straight-identified participants) and lowest among sexual minority participants. Next, we examined the effects of sexual orientation homophily on self-rated mental health for 2SLGBTQ+ and Straight-identified Participants. These results showed that for 2SLGBTQ+ respondents, there was a positive association between self-rated mental health and sexual orientation homophily (B = 0.199, SE = 0.77, p = 0.01). Meanwhile, this effect was not the same for Straight-Identified Respondents (See Figure 3).

Discussion

The evidence summarized above makes it clear that, for better or worse, homophily plays an important role in the structure of our social lives. Indeed, people tend to affiliate with individuals whom they have more common ground with. Sometimes this means that they self-select into relationships with individuals of the same religion, ethnicity, or political group. While these shared experiences can create a basis for meaningful connection, they can also lead to the exclusion of others. As such, there is likely a need to promote cross-group social interactions, which can help break down systemic forms of discrimination, including homophobia and racism. Of course, we should also be mindful of the potential benefits of in-group affiliation – particularly for marginalized communities who may depend on one another for support and resilience in the face of exclusion, discrimination, and violence. Further research is clearly needed to understand the emergence and maintenance of homophily, as well as strategies to promote meaningful forms of diversity. Furthermore, there is need for methodological refinement to understand homophily across multiple identity traits – going beyond simplistic demographic measures. Similarly, it is important to understand the motives and mechanisms that give rise to homophily in order to fully understand its significance and role in our social lives.

Conclusion

Based on the available evidence, we recommend that communities engage thoughtfully with the challenges and opportunities related to the homophily and diversity in our social networks. In particular, efforts to help individuals establish common ground with others can be an important basis upon which friendships can be built. However, it is also important that appropriate steps are taken to promote diversity and inclusion and to provide safe spaces for marginalized individuals. Continued research improving our understanding of the role of homophily and diversity can empower communities in these pursuits.