How attributes of humorous scientific messages predict engagement on Twitter and Instagram

Abstract: The use of humor is increasingly advocated as a means of enhancing the effectiveness and visibility of science messages on social media. However, the influence of humorous scientific content on user engagement is empirically unknown. The contribution of this study is threefold. First, we conduct a content analysis of humorous scientific posts on Twitter and Instagram to shed light on the poster qualities (number of followers and number of accounts following), technical attributes (presence of emojis and visuals, and number of hashtags), and content characteristics (presence of message purposes and humor types). Second, using regression models, we examine how these poster and message attributes are associated with multi-dimensional engagement with the posts in the form of liking, retweeting, and commenting. And third, this study investigates subtypes of humor (e.g., satire, wordplay, and anthropomorphism) embedded in funny scientific messages and their effects on engagement. These findings have implications for science communication practices on social media.