Abstract

Social media moderation encompasses actions undertaken by platforms to uphold community norms and diminish anti-social behaviors (Wise et al., 2006). While existing literature predominantly examines the direct impact of moderation on reducing harmful behavior, limited attention has been given to its spillover effect, particularly the effect on unmoderated content. In this study, we exploit a temporary shutdown of the commenting function on a large Q&A platform, where undisclosed comment moderation was conducted during the shutdown period, to investigate the spillover effect of comment moderation on answer generation. Using a difference-in-differences approach combined with propensity score matching, we find that comment moderation induces a significant decrease in the volume of the subsequent answers, especially for normal answers rather than for harmful answers; Interestingly, the comment moderation improves the helpfulness of the subsequent answers. Our mechanism analysis reveals that the observed effects on answer volume and helpfulness are more pronounced for opinion-oriented, popular questions, compared to those for professional-oriented, niche questions. Further, content analysis shows that moderation leads to longer, more concrete, and more similar answers. We attribute these results to the chilling effect and loss aversion effect among answer contributors on the supply side. On the demand side, moderation reduces users’ content consumption needs, which in turn dampens the answer generation. These findings suggest that well-intentioned moderation may inadvertently erode user trust and participation in content contribution, but improve the helpfulness of user contribution, highlighting trade-offs in platform governance. 

Please contact dip.mkt@unibocconi.it if you wish to attend.