The impact of Facebook's vaccine misinformation policy on user endorsements of vaccine content: An interrupted time series analysis

Objectives: To evaluate the impact of Facebook's vaccine misinformation policy in March 2019 on user endorsements of vaccine content on its platform. Methods: We identified 172 anti-and pro-vaccine Facebook Pages and collected posts from these Pages six months before and after the policy. Using interrupted time series regression models, we evaluated the policy impact on user endorsements (i.e., likes) of anti-and pro-vaccine posts on Facebook. Results: The number of likes for posts on anti-vaccine Pages had decreased after the policy implementa-tion (policy = 153.2, p < 0.05; policy*day =-0.838, p < 0.05; marginal effect at the mean =-22.74, p < 0.01; marginal effect at the median =-24.56, p < 0.01). When the number of subscribers was consid-ered, the policy effect on the number of likes for anti-vaccine posts was much smaller, but still statistically significant (policy = 4.849, p < 0.05; policy*day =-0.027, p < 0.05; marginal effect at the mean =-0.742, p < 0.01; marginal effect at the median =-0.800, p < 0.01). There was no policy effect observed for posts on pro-vaccine Pages. Conclusions: Our analysis suggested that Facebook's March 2019 vaccine misinformation policy moder-ately impacted the number of endorsements of anti-vaccine content on its platform. Social media com-panies can take measures to limit the popularity of anti-vaccine content by reducing their reach and visibility. Future research efforts should focus on evaluating additional policies and examining policies across platforms. (c) 2022 Published by Elsevier Ltd.