As false information may spread rapidly on social media, a profound understanding of how it can be debunked is required. This study offers empirical insights into the development of rumors after they are debunked, the various user groups who are involved in the process, and their network structures. As crisis situations are highly sensitive to the spread of rumors, Twitter posts from during the 2017 G20 summit are examined. Tweets regarding five rumors that were debunked during this event were manually coded into the following categories: rumor, debunking message, uncertainty about rumor, uncertainty about debunking message, and others. Our findings show that rumors which are debunked early and vehemently by official sources are the most likely to be stopped. When individuals participate in the process, they typically do so by sharing uncommented media content, as opposed to contributing user-generated content. Depending on the conditions in which a rumor arises, different network structures can be found. Since some rumors are easier for individuals to verify than others, our results have implications for the priorities of journalists and official sources.

中文翻译：

---

警告：谣言在前——关于在 Twitter 上揭穿虚假信息的案例研究

由于虚假信息可能会在社交媒体上迅速传播，因此需要深入了解如何揭穿它。本研究为谣言被揭穿后的发展、参与该过程的各种用户群体及其网络结构提供了实证见解。由于危机局势对谣言的传播高度敏感，因此审查了 2017 年 G20 峰会期间的 Twitter 帖子。在此事件中被揭穿的五个谣言的推文被手动编码为以下类别：谣言、揭穿消息、谣言的不确定性、揭穿消息的不确定性等。我们的调查结果表明，被官方消息来源及早并强烈揭穿的谣言最有可能被阻止。当个人参与这个过程时，他们通常通过共享未评论的媒体内容来实现这一点，而不是贡献用户生成的内容。根据谣言产生的条件，可以找到不同的网络结构。由于某些谣言对个人而言比其他谣言更容易验证，因此我们的结果对记者和官方消息来源的优先事项有影响。