Abstract
During public health emergencies, as one of the most effective rumor management strategies, rumor rebuttals depend on users' cognition, decision-making and interactive behaviors. Taking the dissemination of rumor rebuttals related to COVID-19 epidemic in the early stage in China as an example, we firstly adapted network analysis to construct representative networks of information and communication flow networks of users based on users' retweeting and commenting behaviors. Then quantitative indicators and exponential random graph models were used to evaluate the level of homophily based on topic and veracity in information networks, identity and standpoint in user networks. Meanwhile, chi square tests were added to compare the degree of echo chamber effect in retweeting and commenting. Findings showed that, users did show significant echo chamber effect when retweeting or commenting on rumor rebuttal information with different veracity. They showed diversification when retweeting but a certain tendency and pertinence when commenting in topic selection. Weibo's direct and open platform for retweeting and commenting broke the boundaries between stakeholders from different professional fields. However, the retweeting mechanism promoted self-isolation of users' standpoints, while the commenting mechanism provided an understanding and integrating channel for groups with opposing standpoints.