Skip to main content

Political Communication and Issue Frames

Networks of Polarization and Unity in U.S. Presidential Primary Campaigns 

Moral rhetoric—or emphasizing ideas of what is right and wrong—is a powerful persuasion technique for candidates running for election. In PNAS Nexus, Kellogg social psychologist and IPR associate William Brady and his colleagues analyze how the use of specific moral words connects or differentiates political candidates in recent U.S. elections. The researchers examined a dataset of 139,412 tweets published by 39 U.S. presidential candidates during the 2016 and 2020 primary elections. They pulled moral language from the tweets, placing the language into five categories: care, fairness, loyalty, authority, and sanctity. The researchers then created network models to reveal patterns and points of distinction in the candidates’ rhetoric. They find that candidates “push away” from the other party with Democratic candidates using more care and fairness language, while Republican candidates used more loyalty, authority, and sanctity language. They also discover that candidates from each party used moral language in highly similar ways, demonstrating how outsider candidates like Donald Trump can separate themselves from others by avoiding the party’s common moral language. The findings show how candidates can use moral language to isolate themselves from their peers, like Trump did in 2016, or to insulate themselves among them, like Joe Biden did in 2020. The researchers suggest that the way candidates express moral rhetoric online may contribute to polarization in social networks. They highlight several future directions for research, including mapping language across election cycles and in different types of elections.