Dipartimento di Economia e Statistica "Cognetti de Martiis"

The way we get, process and share information on social media is characterized by certain unique features. We are exposed to articles and opinions shared by our friends and those who we choose to follow or interact with. When we see an article, we then choose whether to like it ourselves, as a way to signal to others something about our opinions or beliefs. How these distinct yet prevalent features of our online media environment affect how we assimilate information and form personal views, however, is not clear. We propose to study such an information environment both through a theoretic model of information diffusion and attitude formation, and through an online controlled experiment in the context of the migration debate within the EU, an increasingly important and complex social issue.

To measure and understand individuals' susceptibility to information signals, we conduct an online controlled experiment in Italy and Ireland, which aims at measuring the following three features, which are important parameters of our theoretical framework: i) the extent to which users are sensitive to online popularity metrics (i.e. how more likely are users to change their views in response to popular signals?); ii) the extent to which endorsing signals contributes to reinforcing users' views (i.e. do users' conviction increases when they endorse a signal?); iii) the extent to which users respond negatively to opposing views. We also include one policy variation: we warn users about the presence of bots and inflated statistics in online media.

The study aims to contribute to our understanding of online media, political polarization, the impact of fake news and of online bots. By measuring how sensitive users are to our experimental treatments, we aim to understand whether and how bots contribute to political polarization online and citizens' views are malleable on the subject of migration.