Processes of Polarisation across Social Media Platforms

The next speaker in this IAMCR 2019 session is Christian Baden, who shifts our focus to processes of polarisation. Some existing work on polarisation focusses on the themes and content along which groups are polarised, but in itself such differences may not be problematic; rather, the key issue here is whether such polarisation is increasing and results in incompatible perspectives.

Homophily and antagonism drive such processes. But homophily is extremely common and not necessarily a problem in itself; some homophily is natural, and it only becomes a problem in extreme situations. The question is therefore whether homophily increases over time and therefore drives a spiral of mutual exclusion. Similarly, antagonism is also common in politics, and again the more important question is whether it increases over time and reflects political positions.

Such processes are also affected by the shape and structure of the public sphere, and now perhaps especially also by the social media platforms where much public debate now takes place. The project examines this in the context of a recent case in Israel where a soldier killed an already disarmed Palestinian attacker at point-blank range; it drew on an automated analysis of content from Twitter, Facebook, and WhatsApp (and thus also examined wholly public, semi-public, and entirely non-public discussions).

Interactions on Twitter and WhatsApp were largely homophilous, while Facebook interactions were increasingly heterophilous. On Twitter this increased the expression of extreme views, and resulted in gradually more antagonistic responses across polarised groups; this shows some evidenced of attitudinal polarisation. On Facebook and WhatsApp the situation is different, and there is even some evidence of depolarisation in some cases. This means that we need far more sophisticated approaches to understanding the processes of polarisation, that may also need to be platform-specific.