Shame on Who? Experimentally Reducing Shame During Political Arguments on Twitter

Abstract

Online political arguments have a reputation for being futile exchanges, partially because people often respond more punitively to those who do not share their views, a phenomenon called ingroup bias. We explore how ingroup bias affects political disagreements online, and how respect can mitigate its effects. Towards this goal, we conducted an experiment on Twitter systematically varying respectful versus neutral disagreement language across people who did and did not share views. We found that people who do not share views were most likely to reply to disagreements, and neutral disagreements generated more discussions than respectful disagreements. However, we also found that using respectful language increased respectful language received in return, and it reduced the effects of ingroup bias across conversations with people who do and do not share the same views. We conclude with recommendations to promote respectful language on social media and build shame resiliency online, such as designs that encourages thoughtful engagement and a peer support network that allows users to share shame experiences online.