The Spatial Dimension of Online Echo Chambers

Abstract This study explores the geographic dependencies of echo-chamber communication on Twitter during the Brexit referendum campaign. We review the literature on filter bubbles, echo chambers, and polarization to test five hypotheses positing that echo-chamber communication is associated with homophily in the physical world, chiefly the geographic proximity between users advocating sides of the campaign. The results support the hypothesis that echo chambers in the Leave campaign are associated with geographic propinquity, whereas in the Remain campaign the reverse relationship was found. This study presents evidence that geographically proximate social enclaves interact with polarized political discussion where echo-chamber communication is observed. The article concludes with a discussion of these findings and the contribution to research on filter bubbles and echo chambers.

Introduction Literature on online social networks defines echo chambers as a process of self-selection in which communication is circumscribed to ideologically-aligned cliques (Del Vicario et al., 2016; Del Vicario, Zollo, Caldarelli, Scala, & Quattrociocchi, 2017; Zollo et al., 2017). The political communication literature has explored the potential of echo chambers to foreclose deliberation by reinforcing the political views and preferences of individuals (Colleoni, Rozza, & Arvidsson, 2014; Wojcieszak & Mutz, 2009), a development largely seen as a problematic for democracy in as far as it engenders political polarization (Sunstein, 2007). Echo chambers are deeply embedded in processes of political polarization, selective exposure, and “filter bubbles,” but research exploring the structural equivalence (Wasserman & Faust, 1994) of actors involved in

echo-chamber communication remains largely uncharted (Barberá, Jost, Nagler, Tucker, & Bonneau, 2015). While exact equivalence is rarely observed in real-world social networks, structural, isomorphic, and regular equivalence are measures of network similarity and equivalence that can help identifying classes or clusters of users (Freeman, 2011). Transposed to the network of tweets about the U.K. E.U. membership referendum, we seek to explore whether users engaging in echo-chamber communication are clustered in subgraphs that are homogeneous with regard to sociodemographic variables such as household and employment. More specifically, we consider whether geography and class are variables that interact with echo-chamber communication, a proposition in line with the principle of homophily asserting that individuals are more likely to associate and form ties with similar others. Isomorphic dependencies affecting homophily include a range of factors such as age, gender, class, and organizational role (Scott & Carrington, 2011), but for the purposes of this study we sought to extensively probe the relationship between geography and echo chambers.

Homophily thus offers a framework for understanding isomorphic dependencies that might interact with echo chamber communication. Firstly, echo chambers censor, disallow, or underrepresent competing views by enforcing social homogeneity much in line with the bandwagon effect predicted by the homophily model (Mark, 2003). Secondly, cultural similarities and differences among people can be formalized as a function of geographic propinquity (McPherson, Smith-Lovin, & Cook, 2001). Thirdly, online social networks are more prone to homophily compared with offline networks, which are tied to physical locations where serendipitous exposure to social diversity is more likely to happen (Hampton & Gupta, 2008; Hampton, Livio, & Goulet, 2010). In view of that, we seek to advance cognate scholarship with an investigation into whether echo-chamber communication is associated with geographical proximity between users that tweeted the U.K. E.U. membership referendum or the “Brexit campaign.” The Brexit referendum campaign was held at a time of sharp polarization among the electorate on the cultural and economic costs or, alternatively, the benefits of E.U. membership (Kriesi & Frey, 2008). While older and culturally conservative voters protested the infringement on national sovereignty by European institutions, in addition to being concerned about the impact of E.U. workforce mobility on British society, a more liberal-minded electorate welcomed the

collective decision-making at the heart of the European Union and accepted the inward E.U. migration into the U.K. as a source of greater diversity (Curtice, 2016). In the run-up to the referendum, the winning campaign to leave the Union foregrounded a culturally conservative message centered on the proposition to ‘take back control’

…(Full text truncated)…