My second day at the IAMCR 2025 conference in Singapore starts with a session on polarisation and media effects, which begins with a paper by Rachel Neo and Benjamin Johnson, whose focus is on affective polarisation on Reddit in the US. The US is now deeply polarised, and this also expresses in incivility from both sides of politics. Such incivility is visible in disrespectful exchanges between partisans, which is often expressed affectively.
This is a case of affective polarisation: positive feelings towards the in-group, and negative feelings towards the out-group. Reddit is one space online where this can be seen; the US has some 48 million daily active Redditors, who may sort themselves into politically broadly homophilous subreddits. But such partisan sorting is far from complete: incivility still occurs there, even in spite of the presence of moderators.
This study examines how this might be mediated by several factors: in-party prototypicality (perceptions on how much the information encountered represents one’s own party’s views), out-party prototypicality, perceived comment manipulation by the out-party, and perceived intolerance towards opposing political views. To test this, the study conducted a survey experiment, exploring the impact of comment valence and comment volume.
Participants were shown one thread from a Democrats or Republican subreddit, featuring uncivil language towards one or the other side of US politics. Participants were then asked for their perceptions of the proposed mediators. Attitude valence and comment volumes had medium to strong effects.
Pro-attitudinal comments affect out-party negative affect and in-party positive affect either directly or via in-party prototypicality. Perceptions of comment manipulation and perceived intolerance also influenced both forms of affect. Comments favouring the other party are more likely to be perceived as manipulated and intolerant. There was no total affect of comment valence on affective polarisation, though.
This is broadly consistent with past research on affective polarisation: counter-attitudinal information increases positive feelings towards one’s own party, and perceptions of comment manipulation; and also increases perceptions of in-party prototypicality. But the indirect effects of comments on hostile feelings towards the out-party was more surprising: it is possible that pro-attitudinal information reminds people to be hostile towards the out-party, and to suspect manipulation.
Comment volume on Reddit seemed to have little effect; other Reddit cues may be more important, though (such as clearly partisan usernames and avatars, etc.).
