Depolarization in the Rise of Far-Right Platforms? A Moderated Mediation Model on Political Identity, Misinformation Belief and Voting Behavior in the 2020 US Presidential Election

Introduction. Featured with full anonymity and image-based communication, imageboard has created a special online public space, well known for the widespread cultural and political memes (Nissenbaum & Shifman, 2017) as well as the far-right and populists’ antagonistic speech (Nagle, 2017; Tuters & Hagen, 2019). While empirical analysis on the contents gradually cumulates (Bernstein et al., 2011), little is known about the political implications of using these media platforms. This paper thus attempts to explore the topic by examining how the use of imageboard conditions the impact of political identity on people’s misinformation belief and the downstream effect on voting behavior in the 2020 United States Presidential Election. Methods. With a national representative survey (N = 702), we run regression models on two mediators (belief in con-Trump and con-Biden misinformation) and one binary outcome variable (0 = vote for Biden, 1 = vote for Trump), using political identity as the key explanatory variable (0 = extremely liberal, 10 = extremely conservative) and general Internet use, news consumption on all media, and demographics as controls. Further, we take respondents’ experiences (0 = used, 1 = unused) in using the two most popular imageboards, 4chan and 8chan (8kun), as a moderator in the relationships between political identity and the two mediators. Results. Political identity is highly correlated to belief in both con-Trump and con-Biden  misinformation and they all significantly predict the eventual voting behavior. More interestingly, imageboard use presents a significant moderating effect on the relationships between political identity, the two types of misbelief (belief in con-Trump misinformation: B = 0.10, p < 0.05; belief in con-Biden misinformation, B = -0.17, p < 0.001), and voting behavior (B = -0.31, p < 0.001). For imageboard users, the effects of political identity on misinformation belief as well as voting behavior are much more mitigated, i.e., closer to zero, than those of non-users. With a bootstrapping approach (Preacher & Hayes, 2004), the moderated mediation model is tested significant only for the indirect effect of the interaction via belief in con-Biden misinformation. Discussion. Previous studies on imageboards suggest that contents on those platforms are closely connected to far-right political movements (Ylä-Anttila et al., 2020). Counter-intuitively, our findings suggest that the use of imageboards mitigates partisans’ existing political identity-based misbelief about the presidential candidate, with downstream consequences in voting behavior. One possible explanation could be that exposure to the extreme (possibly indecent) alt-right contents on the imageboards leads to people’s reflections on the party leaders (for both the liberals and the conservatives), suggesting a potential “depolarization effect”. This result coincides with a recent observation of a split between the traditional ring-wing and the far-right, e.g. the QAnon activists (Tollefson, 2021). Future work could incorporate more online data through web crawling to further examine the connections between online participation in imageboards and offline political consequences.