Talk, CSSLab@CUHK, Hong Kong
(Normative) theorists have argued that conversation is the soul of democracy. However, political discussions might not be as ideal as the theorists envisioned. This talk summarizes a series of computational studies using large-scale social media data to examine the structures and consequences of political discussions online. Specifically, this talk discusses both the positive and negative sides: the organizational principles of web forum discussions and how they could help produce common ground among the users, how diffusion structures are related to cross-ideological interaction on Twitter, to what extent movement-related discussions on LIHKG are leaderless, and finally how to identify the causal effects of incivility such as swearing and name-calling using machine learning and text mining.