Latest research 2026: are comment sections harmful?

The strongest evidence shows that active moderation with visible norms and graduated enforcement improves civility, while hiding like counts did little for well-being and switching to a chronological feed did not reduce affective polarisation. The most common response to hostile comments is withdrawal or exit, not attitude change, so studies that only measure the people who stay likely overstate polarisation. Users widely describe a compulsive 'debate trap' with fight-or-flight arousal and rumination that standard well-being scales may miss, but this rests on self-reports rather than controlled studies. The link between heavy commenting and poor well-being is heavily confounded by personality and survivor bias, so causal claims should be treated with caution.

Main caveats: Causal evidence at the comment-thread level is lacking; most findings come from feed-level experiments or lab studies. Publication bias inflates the appearance of uniform effects, and the field's evidence base is skewed toward Twitter/X and political contexts.