Examining spread of emotional political content among Democratic and Republican candidates during the 2018 US mid-term elections

Previous research investigating the transmission of political messaging has primarily taken a valence-based approach leaving it unclear how specific emotions influence the spread of candidates' messages, particularly in a social media context. Moreover, such work does not examine if any differences exist across major political parties (i.e., Democrats vs. Republicans) in their responses to each type of emotional content. Leveraging more than 7000 original messages published by Senate candidates on Twitter leading up to the 2018 US mid-term elections, the present study utilizes an advanced natural language tool (i.e., IBM Tone Analyzer) to examine how candidates' multidimensional discrete emotions (i.e., joy, anger, fear, sadness, and confidence) displayed in a given tweet—might be more likely to garner the public's attention online. While the results indicate that positive joy-signaling tweets are less likely to be retweeted or favorited on both sides of the political spectrum, the presence of anger- and fear-signaling tweets were significantly associated with increased diffusion among Republican and Democrat networks, respectively. Neither expressions of confidence nor sadness had an impact on retweet or favorite counts. Given the ubiquity of social media in contemporary politics, here we provide a starting point from which to disentangle the role of specific emotions in the proliferation of political messages, shedding light on the ways in which political candidates gain potential exposure throughout the election cycle.

Meng-Jie Wang, Kumar Yogeeswaran, Sivanand Sivaram, Kyle Nash