Drivers of True and False Information Spread: A Causal Study of User Sharing Behaviors

Analyzing and predicting user information-sharing behavior on online social platforms is a crucial task in social sciences. While current prediction tasks primarily emphasize accuracy, they often neglect the underlying motivations that drive user behavior, hindering a fundamental understanding and control of the information spreading environment. To address this, we analyze and quantify potential factors that may drive user sharing behavior based on social theories. Our limited derived feature set achieves over 85% accuracy in predicting user behavior on two real-world datasets, demonstrating its effectiveness. Notably, through employing causal inference techniques, our analysis on true and false information spread reveals that users with lower authority are more susceptible to being misled by false information. In contrast, the propagation of truthful news is often driven by personal preference or influenced by users’ social circles. By uncovering these underlying motivations, our approach facilitates a deeper comprehension of the online information ecosystem, contributing to more effective management strategies for false information mitigation.