The Emergence of Polarized Social Networks via Information Cascades
Research Question: How exactly does partisan media coverage polarize society? Many argue that partisan media coverage causes polarization by pushing people's opinions to the extreme, but evidence is mixed. We instead propose that partisan media coverage might polarize society by altering people's social connections and reorganizing social networks along political lines, i.e., create so-called "echo chambers."
Data: Our empirical data was sourced from Twitter's API. We also attempted some experimental methods on Twitter and recruited participants through MTurk and Facebook Ads.
Methods: We first built a theoretical agent-based model that explores how people may adjust their social ties to avoid the sharing behavior of friends who might be engaging with news from nonpreferred information sources. We then tested the predictions of this model through observational data collection using Twitter's API. While not included in the final paper, we also attempted to create experimental social networks (treated by media environment) on Twitter, and we recruited participants through MTurk and Facebook Ad campaigns.
Challenges: Rate limits on Twitter's API; Recruitment of experiment participants; Matching up experiment participants with social media handles while adhering to platform regulations.
Findings: We show that polarized media coverage—that is, when news sources assign very different importance to the same topic—creates information cascades that cause "echo chambers" to form in social networks, even when people do not know each other's political identity. These echo chambers also limit the flow of information, causing people in them to miss out on news from both their preferred and non-preferred outlets. We verified these predictions on Twitter, finding that people who follow more polarized media outlets lose opposite-ideology at a faster rate than expected by chance, suggesting that they are sorting into politically homogeneous social networks.