Does Scientific Evidence Sell? Combining Manual and Automated Content Analysis to Investigate Scientists’ and Laypeople’s Evidence Practices on Social Media

Abstract

Examining the dissemination of evidence on social media, we analyzed the discourse around eight visible scientists in the context of COVID-19. Using manual ( N = 1,406) and automated coding ( N = 42,640) on an account-based tracked Twitter/X dataset capturing scientists’ activities and eliciting reactions over six 2-week periods, we found that visible scientists’ tweets included more scientific evidence. However, public reactions contained more anecdotal evidence. Findings indicate that evidence can be a message characteristic leading to greater tweet dissemination. Implications for scientists, including explicitly incorporating scientific evidence in their communication and examining evidence in science communication research, are discussed.