Unraveling the chilling-backfire dynamic: A mixed-methods study on misinformation moderation through the lens of spreaders

In response to the rampant spread of misinformation, social media platforms are increasingly deploying content moderation measures. Among them, warning labels have garnered significant attention. Prior research has predominantly examined labeling effect from receiver side, overlooking the role of spreaders who initiate and disseminate the false content. Moreover, evidence regarding labeling effectiveness remains inconsistent, showing both chilling and backfire effects, while often relying on self-report data that may not capture actual long-term behavior. We adopt a mixed research method, combining unobtrusive data and in-depth interviews, to investigate how warning labels influence misinformation spreaders’ posting frequency as well as their perceptions and responses. Analysis of 67,286 Weibo platform data reveals a backfire effect of misinformation labels, where, labeled spreaders significantly increased their posting frequency. In-depth interviews further confirmed this finding, revealing that warning labels are often perceived as opaque instruments of control, which, in turn, triggered a rebellious mindset among users.