Linking User Opinion Dynamics and Online Discussions

This paper studies the dynamics of opinion formation and polarization in social media. We investigate whether the stance of users with respect to contentious subjects is influenced by the online discussions that they are exposed to, and by the interactions with users supporting different stances. We set up a series of predictive exercises, in which users are described using several posting activity features capturing their overall activity levels, their posting success, the reactions their posts attract from users of different stances, and the types of discussions in which they engage. Given the user description at the present time, the purpose is to predict his stance in the future. On a dataset of discussions about Brexit on the Reddit platform, we show that the activity features constantly outperform the textual baseline, confirming the link between exposure to discussion and opinion. We find that the most informative features relay to the stance composition of the discussion that users prefer to engage into.