Social Media Governance: Disinformation Detection, User Moderation, and Newsfeed Regulation

Social media platforms can profoundly influence people’s beliefs and behaviors, with significant social and economic implications. However, the presence of inappropriate content on these platforms—including but not limited to misinformation, cyberbullying, and hate speech—can counteract potential benefits, leading to non-negligible economic losses and negative societal consequences. It becomes imperative for social media platforms to implement effective governance strategies to control harmful content, moderate inappropriate users, and promote beneficial engagements. In my dissertation, I design IT artifacts and empirically evaluate the effectiveness of diverse strategies pertaining to social media governance. In the first study, I develop a well-justified machine learning system to detect financial disinformation published on social media platforms. Extensive analyses are conducted to evaluate the performance and efficacy of the proposed system. In the second study, I focus on user ban, a common but controversial moderation strategy. Specifically, I investigate the impacts of user bans on banned users’ content-generating behaviors (both quantity and quality). In the third study, I investigate the relationships across newsfeed channels on social media: social networks, algorithmic recommendations, and trending content. Prior literature has investigated the impacts on user-content engagement within each channel. However, little is known about the relationships across these channels. In particular, I study how limiting content display from the social network channel can influence the quantity and the diversity of user-engaged content across channels. The three studies offer various theoretical implications and practical values.