Three studies into user engagement behaviors on social media : the dynamics of fake news and rumor-debunking

Abstract

This thesis investigates user engagement behaviors on social media platforms affected by fake news dissemination through three studies. These three studies collectively enhance our understanding of user engagement behaviors on social media across two dimensions. First, they elucidate the influence process of fake news on user engagement behaviors, from initial impact to the effects of subsequent interventions. The first two studies specifically investigate how news related factors impacts user engagement, while the third study delves into the effects of rumor-debunking content on user engagement behaviors. Second, the studies investigate the factors influencing user engagement behaviors by analyzing both the overall environment and detailed features. The first study explores how overall news quality on social media influences user engagement behaviors, while the second and third studies examine how specific features of online news and rumor-debunking content impact user engagement behaviors. Study 1 utilized the Social Amplification of Risk Framework and Structural Equation Modeling to assess how news quality influences users’ risk perceptions, perceived believability, and news-sharing behaviors. Additionally, I explored the moderating effects of fake news awareness and social tie variety. The findings indicated that the influence of news quality on users’ news-sharing behavior is mediated by risk perception and perceived believability. Individuals with a heightened awareness of fake news or a diverse social tie are more inclined to perceive greater risks associated with news-sharing behavior and question news authenticity. Study 2 applied Media Richness Theory and negative binomial regression model to investigate how textual and pictural information richness influence user engagement. I also explored the moderating effects of the type of discussed travel destination (natural or cultural landscape), complexity of discussed travel destination, and poster level in social media (common or VIP), using data crawled from Weibo platform. The results showed that textual information richness exhibits a significantly positive influence on user engagement and pictorial information richness demonstrates a significantly negative effect on user engagement. The complexity of discussed travel destination and poster level positively moderate the impact of both pictural and textual information richness on user engagement. Additionally, when the topic under discussion is the natural landscape, both the positive influence of textual information richness and the negative influence of pictural information richness will decrease. Study 3 draws on theories of Impression Management and Source Credibility, conducting a polit study and three experiments to investigate how fact-checkers affect user engagement with rumor-debunking content. The study identifies a dual mediation process involving innovation image and perceived source credibility, moderated by AI literacy and confirmation bias. Fact-checkers indirectly boost engagement through these mediators, with AI literacy and confirmation bias altering the mediation paths. The results indicated a significant indirect effect of fact checkers on user engagement through the mediating roles of innovation image and perceived source credibility. AI literacy primarily moderates the path of innovation image, while confirmation bias moderates the path involving perceived source credibility. These findings contribute to the literature on user engagement with news and AI-generated rumor debunking, offering practical insights for curbing fake news spread and enhancing fact-checking effectiveness.