Twitter and Facebook posts about COVID-19 are less likely to spread false and low-credibility content compared to other health topics (2007.09682v3)

Abstract: On February 2, 2020, the World Health Organization declared a COVID-19 social media "infodemic", with special attention to misinformation -- frequently understood as false claims. To understand the infodemic's scope and scale, we analyzed over 500 million posts from Twitter and Facebook about COVID-19 and other health topics, between March 8 and May 1, 2020. Following prior work, we assumed URL source credibility is a proxy for false content, but we also tested this assumption. Contrary to expectations, we found that messages about COVID-19 were more likely to contain links to more credible sources. Additionally, messages linking to government sources, and to news with intermediate credibility, were shared more often, on average, than links to non-credible sources. These results suggest that more ambiguous forms of misinformation about COVID-19 may be more likely to be disseminated through credible sources when compared to other health topics. Furthermore, the assumption that credibility is an adequate proxy for false content may overestimate the prevalence of false content online: less than 25% of posts linking to the least credible sources contained false content. Our results emphasize the importance of distinguishing between explicit falsehoods and more ambiguous forms of misinformation due to the search for meaning in an environment of scientific uncertainty.