UC Davis

Discourse incivility is rampant online and has been linked to various adverse democratic outcomes, including political polarization, extremist attitudes, and misinformation. This work centers around understanding the affordances of social media that can disincentivize uncivil discourse in both political and nonpolitical contexts.Reports of sexual assault have been found to elicit online discourse incivility. Chapter 2 employed a computerized coding tool to examine linguistic characteristics of news media that are likely to influence discourse incivility—specifically, negative emotion, disagreement, and discussion about power relations. Additionally, machine learning was harnessed to measure the levels of comment toxicity, insult, profanity, threat, and identity attack in Reddit and Twitter posts sharing news reports of sexual assault. Findings revealed that linguistic features of news articles interact with platform community norms to predict rape culture as expressed within online responses to reports of sexual assault. Fanship identities are an important part of some individual's self-concept. Chapter 3 investigated incivility as it relates to the perceived presence of similar others—specifically in the context of sports fanship, which is a situationally-salient identity. While much of the sports fan literature has investigated incivility between fans of opposing teams, we investigated incivility on fan forums dedicated to specific teams—specifically the levels of incivility in each of the NFL fan subreddits as it relates to the perceived presence of similar others (i.e., the number of followers of each subreddit total number of households in the media market each team is located). Results revealed a negative relationship between the total size of fandom (media market) and 3/5 dimensions of incivility—toxicity, insults, and profanity. Findings suggest that--in the presence of similar others, incivility declined with a larger perceived audience. In other words, the more folks users perceive share their fan identity when talking to an audience comprised of fellow fans, the less likely users are to be uncivil. Issues of identity and online behaviors are closely linked; individuals’ self-esteem needs are shaped--in part--by their sense of group inclusion and the perceived value of the groups we belong to. Political partisanship is a social identity that causes this sense of self and ‘us’ vs. ‘them’ attitude, manifesting as uncivil behavior toward the political outgroup online. Yet, as social beings, individuals adapt their communication to the situation at hand — the "context." Chapter 4 employed an interactive mock social media platform to test whether users act more civilly toward political opposition in cross-cutting discourse about a polarized health topic (i.e., making birth control easily accessible and free to all college students) when they share nonpolitical identities with them—or feel like they are more similar. Likewise, we tested whether civil conversation can lead to attitude convergence rather than reinforcement, as well as reduced affective polarization. While findings did not show changes in political polarization, results showed that when individuals share nonpolitical identities with political outgroup members in cross-cutting exchanges, they act more civilly than when they do not. Overall, these studies demonstrate how the interactions between social media platform features, user identities, and perceived audience identities can be leveraged to promote civil discourse in three contexts and extend the use of linguistic methodologies to measure incivility in naturalistic and experimental settings. Implications and future directions are discussed.