Tell Us How You Feel: Emotional Appeals for Votes in Presidential Primaries

• In Event: Pretty Persuasion: Campaign Communication in a Social Media Era

Abstract

A growing body of literature traces the powerful effects of emotional cues on attitudes and political behavioral, especially within electoral contexts. Given these powerful effects, understanding patterns of the utilization of emotional cues by political elites is of the utmost importance. Yet we know little about the usage of emotional cues, especially in primaries where studies of candidate rhetoric are still overwhelmingly dedicated to other topics like attacks, issue agendas, and populism. Studying emotional cues in primaries is particularly important as the patterns of emotional cues may in turn affect the party’s electorate in the general election. If one party tends to use more emotional language, then that party’s electorate may begin the general election season more emotionally charged. In this paper, we study this topic by examining differences by party in emotional cue language in speeches by presidential primary candidates. Drawing on theories in the study of personality and social cognition, we hypothesize that Republican candidates will use more emotional language, on average, than Democratic candidates. As extant research suggests, self-identified conservatives are more likely to be intolerant of ambiguity and desire closure, creating an incentive for Republican candidates to make more emotional appeals. Leveraging appraisal theory, we furthermore provide a rationale for why Republican candidates (namely those who are behind in the polls) so often use the language of both fear and joy in public speeches, seeking to motivate would-be voters to consider all their options. We test these hypotheses using the Presidential Primary Communication Corpus which contains more than 2700 speeches by primary candidates from 2000-2016. We measure the amount of emotional language primarily by applying the EmoLex dictionaries of emotional language to the corpus with replication analysis using the LIWC dictionaries. We supplement this analysis with public polling data of self-reported emotions and attitudes toward presidential candidates. Preliminary results suggest that Republicans use more emotional language than Democratic candidates, as hypothesized, and that this is true of both positive and negative emotional language. This research offers a significant contribution to the study of political communication in presidential primaries by highlighting how the reactions of partisans to emotional appeals powerfully shapes the rhetorical choices of candidates.

Authors

• Zachary Scott, Wake Forest University

• Jared McDonald, Stanford University