Homophily and Online Politics

Introduction

Homophily is “the principle that contact between similar people occurs at a higher rate than among dissimilar people” (McPherson, Smith-Lovin & Cook 2001, 416). Studies of homophily have a long and distinguished tradition in anthropology, sociology, political science and mass communications. While numerous studies on social networks were published in the 1920s, 1930’s and 1940’s (e.g., Almack 1922, Richardson 1940, Bott 1929, Loomis 1946), the classic citation in the homophily literature is Lazarsfeld and Merton's (1954) analysis of friendship patterns in two small New Jersey towns. In this study, Lazarsfeld and Merton distinguish between status homophily, in which “similarity is based on informal, formal, or ascribed status,” and value homophily, which is based on “values, attitudes, ands beliefs.” Subsequent empirical research into homophilous sorting – the propensity of individuals who are similar on some meaningful dimension to form clusters with each other – has employed this basic distinction as a guide.

Research over the last fifty years has shown that both status and value homophily are driven by many factors. One particularly important factor in the early literature on homophily was geography and physical proximity. As McPherson et al. (2001) write, “it takes more energy to connect to those far away than those who are readily available” (429). Based on this line of work, one might expect that technological innovations that “loosen the bounds of geography by lowering the effort involved in contact” (Kauffer & Carley, 1993) will dramatically minimize both status and value homophily. In particular, the Internet, with its ability to allow asynchronous and spatially unconstrained communication between previously isolated individuals, should diversify the composition of social networks and encourage cross-ideological political discussions.

It is easy to imagine, however, that the Internet will exert a much different kind of impact on homophily. Specifically, in the political realm, the Internet may promote high levels of homophilous sorting along ideological lines. As Farrell (2012) points out, there are a variety of ways in which the Internet makes it more likely that individuals with shared political views will cluster together. According to Farrell, the Internet encourages homophily by making it exceedingly easy for individuals to express their opinions and begin directly interacting with others who share those views. But, as Farrell points out, homophily may also occur more indirectly. Individuals may, for example, spontaneously converge around a common source of online political information that is attractive given their shared interests and cluster together only as a secondary consequence of this shared interest. Thus, while technological change is often perceived as facilitating diverse political interactions, “new technologies may have [just] allowed people greater latitude to create ties that are homophilous” (Hampton & Wellman, 2000).