Asymmetric Dynamics of Partisan Warriors in YouTube Comments

Abstract

Cross-cutting commenting on social media is often imagined as a path to deliberation, yet exposure to opposing views frequently fuels hostility. To explain this dynamic, we introduce the concept of partisan warriors--commenters who cross ideological lines primarily to launch uncivil attacks against out-partisans. We analyze a large corpus of YouTube comments (N= 1,854,320) surrounding the 2024 U.S. second presidential debate. After filtering for toxicity and active participation, we use large language models to identify attack targets and operationalize partisan warrior behavior. Our analysis highlights four dynamics. First, cross-cutting commenters do not exhibit greater civility than those who remain within their ideological camps (RQ1). Second, audience reactions diverge by ideology: conservative audiences tended to reward hostile attacks on out-group leaders, whereas liberal audiences offered no comparable incentives and at times penalized such attacks (RQ2). Third, partisan warriors are notably more prevalent in conservative-leaning channels than in liberal ones; commenters restricted to conservative spaces were substantially more likely to engage in partisan warrior behavior compared to their liberal-only counterparts (RQ3). Finally, regarding environmental triggers, robustness checks suggest that this participation is an ecological phenomenon driven largely by channel-level heterogeneity rather than transient responses to individual video titles (RQ4). By shifting attention from the prevalence of incivility to its targets, rewards, and structural drivers, this study advances understanding of how partisan hostility is enacted and sustained in online spaces.

Figures (8)

• Figure 1: Distribution of toxicity scores by commenter group. Violin plots show the full distribution, with boxplots overlaid for medians and interquartile ranges. Cross-cutting commenters are active on both liberal- and conservative-leaning channels, while only-liberal and only-conservative commenters denote activity restricted to liberal- or conservative-leaning channels. While the overall test indicates significant differences across groups, post-hoc comparisons show no significant difference in toxicity between cross-cutting and only-liberal commenters.
• Figure 2: Incidence Rate Ratios of audience likes for uncivil comments across targets and channel leanings, estimated via negative binomial regression. Error bars indicate 95% confidence intervals.
• Figure 3: Predicted probabilities of PW participation by video title toxicity and channel leaning, based on the Baseline Model. Shaded regions represent 95% confidence intervals. The plot illustrates the apparent ideological asymmetry: toxic titles appear to deter PWs on liberal channels (blue) but have a negligible or slightly mobilizing effect on conservative channels (red).
• Figure 4: Comparison of fixed effects between the Main Model (Baseline, Blue) and the Robust Model (Red). Error bars represent 95% confidence intervals. Notably, the interaction term is significant in the Baseline Model but becomes non-significant in the Robust Model (indicated by the hollow circle), confirming that the apparent asymmetry is driven by channel-level heterogeneity.
• Figure 5: Distribution of $\tau$ for All Users. The overall population exhibits a bimodal distribution, justifying categorical separation.