Democratic and Republican presidential campaigns use fundamentally different moral rhetoric, study finds

July 19, 2023

New research by NOMIS Awardee Manos Tsakiris and colleagues reveals how candidates in US presidential primary elections use distinct moral rhetoric on social media to appeal to voters. Their findings were published in PNAS Nexus.

According to a new study, led by researchers at the Centre for the Politics of Feelings and published in PNAS Nexus, the differences in moral rhetoric used by Republicans and Democrats were so extreme that the researchers were able to identify the party of 34 candidates who ran in the 2016 and 2020 US Presidential primaries solely by mapping their moral language from Twitter. This groundbreaking visualization of moral rhetoric not only reveals candidates’ party affiliations but also allows us to identify instances where individuals like Trump set themselves apart from their fellow primary contenders, as was the case in 2016.

Manos Tsakiris (Photo: Ines Njers)

The study, led by Kobi Hackenburg, a PhD student at the University of Oxford and lab-led by Professor Manos Tsakiris from the Centre for the Politics of Feelings, Royal Holloway, University of London, was carried out in partnership with Professor William J. Brady at Northwestern University.

Using a complete dataset of all tweets posted by 34 candidates throughout the 2016 and 2020 U.S. presidential primaries, they developed a novel methodology combining natural language processing and network analysis to map the moral language used by candidates and organize their campaigns in rhetorical space.

The study focused on five moral values — care, fairness, loyalty, authority, and sanctity — and examined how a candidate’s use of words associated with these values connected or distinguished them from other candidates. The different moral language used was a clear indicator of a candidate’s political affiliation, with the results showing a stark difference between Democrat and Republican candidates.

Democratic and Republican candidates in both 2016 and 2020 primaries could be easily distinguished from one another – not just through their policy language, but also through the moral language they used to argue for the rightness or wrongness of those positions. The two parties emphasized different moral values, with Democrats emphasizing caring and fair treatment of individuals and Republicans emphasizing in-group loyalty and respect for social hierarchies. Moreover, within each party, candidates discussed popular moral values in highly similar ways. For instance, Kamala Harris and Amy Klobuchar used a “care” vocabulary shared by nearly all of their Democratic peers.

(A) Network displaying the moral-rhetorical community structure of the 2016 U.S. presidential primaries, based on a frequency analysis of 574 moral terms used by 17 Democratic and Republican candidates on Twitter. Candidates are connected to each other through their use of the same moral words. (B) Network displaying the moral-rhetorical community structure across both the 2016 and 2020 U.S. presidential primaries, based on a frequency analysis of 574 moral terms used by 34 Democratic and Republican candidates on Twitter. Candidate nodes are connected to each other through their use of the same moral words and colored by partisan affiliation. (Figure: Tsakiris et al. 2023)

The results also reveal the extent and manner of candidate deviations from party norms: for example, Tulsi Gabbard and Andrew Yang deviated towards the Republican network position by using large amounts of conservative loyalty and sanctity language. Conversely, Pete Buttigieg and Joe Biden, Democrats notable for their broad popularity during the primary, also used larger proportions of loyalty and sanctity language but did so in a way which allowed them to retain central network positions within the community of Democratic candidates. In other words, Buttigieg and Biden—even as they used opposing (and more broadly persuasive) moral frameworks—rhetorically insulated themselves amongst their peers by creating new vocabularies of “Democratic” loyalty and sanctity words.

Kobi Hackenburg, from the Oxford Internet Institute at the University of Oxford, said: “Through the deliberate emphasis of certain moral values over others, competing candidates align or contrast themselves with their fellow party members. Our network-based methodology allows us to illustrate these intricate associations, revealing how moral language is associated with party dynamics, ideological shifts, and potentially even electoral outcomes.”

Professor Manos Tsakiris, Professor of Psychology at Royal Holloway, University of London, said: “These patterns of moral expression reveal another dimension of polarization in the US. We illustrate a meaningful divergence in discourse between candidates and parties, and this has major implications for the ways voters engage with campaign messaging, respond to campaign issues, and form opinions about political candidates.”

As we approach the upcoming 2024 U.S. presidential primary season, this research provides fresh insight into the underlying significance of political messaging. When appealing to their respective bases, candidates employ distinct moral frameworks associated with their party affiliations. This consistent divergence in moral perspective has served as a defining characteristic in the electoral landscape of American presidential campaigns in both 2016 and 2020. With the Republican Party already embroiled in a primary battle, this study holds immense value in helping voters and political analysts understand the intricate moral and rhetorical dynamics shaping a fragmented and expansive national debate.

Go to this Royal Holloway release

Read the PNAS Nexus publication: Mapping moral language on US presidential primary campaigns reveals rhetorical networks of political division and unity