Taegyoon Kim (Pennsylvania State University)
Abstract: Although social media contribute to political participation by enabling citizens to freely express and exchange political opinions, increasing concerns are raised about the violent nature of online political communication. Media reports show that social media platforms are exploited by users who write posts threatening, endorsing, or inciting political violence as part of normalized expression of extreme partisan views. Although a small fraction of such violent posts materialize into physical violence, not only do they generate fear on the part of those who are targeted but also they can create a combative climate in online political discussion. Therefore, it is crucial to understand how prevalent such content is and how far it spreads on social media networks. Focusing on Twitter, I suggest an approach to automatically detecting content promoting political violence using machine learning and natural language processing. Once Tweets promoting political violence are identified, I set out to investigate the communication networks through which such violent Tweets spread to potentially large number of users. The preliminary results suggest that a small number of Tweets promoting political violence can cascade into a large size of audience, that such Tweets are more prevalent in conservative subnetworks, and that their spread occurs more frequently among ideologically homogeneous nodes.