Jing Zeng, Queensland University of Technology
Chung-hong Chan, University of Hong Kong (Presenter)
King-wa Fu, University of Hong Kong
This article studies the content moderation strategies used by Weibo to regulate rumour discussion in the wake of the 2015 Tianjin blasts. Over 100,000 Weibo posts were collected and categorised into three datasets, these being rumour discussion posts from the public, rumour-debunking posts from Weibo’s official rumour rebuttal accounts, and posts removed from the system. We have identified two content moderation strategies on rumours, namely, rumour rebuttal and content removal. Clustering analysis and time series analysis were applied to test how these two strategies were used on different topics and how they influenced public discussion on rumour-related topics. Our findings suggest that this platform responds to rumours differently, depending upon the political sensitivity of the topic. We found that Weibo messages were refuted and censored differently based on the political sensitivity of the topics. Time series analysis identified rumour rebuttals and censorship usually associated with more general discussion about the rumour and therefore we have no evidence to support these strategies can consistently create a chilling effect.