PhD defence François Thomas t’Serstevens

Keywords: Fake News, Polarisation, Social Media, Bayesian inference, Computational Social Science

"Truth Perception on Social Media: Mechanisms and Dynamics of Polarisation"

This thesis investigates how people perceive, share, and respond to political information on social media, and how these behaviours can contribute to polarisation. It examines why people are more likely to share information that matches their existing beliefs, particularly among those who share fake news. The findings suggest that what is considered “fake news” depends strongly on who is making the judgement: in the United States, Democrats and Republicans can view each other as sources of fake news even when they do not know the political origin of the information. In Sweden, the research finds that politicians can gain more attention by attacking their political opponents, making negative and polarising messages particularly effective online. Finally, the research examines YouTube and finds that news videos are more likely to become popular when they successfully keep viewers engaged. Together, these findings show how social media can amplify existing beliefs, political conflict, and polarisation.

Click here for the live stream.