From followers to close ties: homophily in population level social networks

The network of social relationships connecting us in society acts as a window through which we view the world, shaping our perceptions and opinions about the social reality surrounding each of us. Therefore, homophily, or the tendency of ties to connect similar individuals, has important consequences for societal cohesion, integration, and polarization. This dissertation focuses on homophily by demographic and political variables for country-level populations, among different types of ties — online and offline, strong and weak. The first chapter studies homophily by age, sex, race, and partisanship in the follower ties of US Twitter users and how geography shapes follower patterns and interacts with homophily, leveraging a dataset of Twitter accounts matched to US voter records. The second chapter examines the relationship between location of residence and social segregation by race among strong ties in the US, using ego-network survey data. The third chapter turns to the association between tie strength and similarity by SES, ideology, gender, country of origin, and religion, drawing on survey data from the Netherlands, Hungary, Poland, and Sweden.