Angry, Abrasive and Unyielding: Gender, Race, Partisanship and the Dynamics of Candidate Likability in the United States

What does it mean for a political candidate to be ‘likable?’ Most models of vote choice in the United States include some element of likability, polls often include questions about how likable a candidate is, and much media attention during a campaign is devoted to the likability (or lack thereof) of the individuals in the race. Yet, as a concept, likability is difficult to define and little political science literature has systematically considered the factors that contribute to or detract from perceptions of likability. What is clear is that candidate gender, race and partisanship all play an important role in determining perceptions of candidate likability, often to the detriment of women, and especially women of color. While women’s political representation has increased substantially, enormous disparities still persist. Perceptions of candidate likability influenced by stereotypes and biases based in gender, race and partisanship may be a key factor in explaining these disparities. In this paper, we examine the relationship between gender, race, partisanship and likability for political candidates. We seek to determine what makes a candidate “likable” or not and whether this differs by candidate gender, race, ethnicity and/or political party. We also consider whether there are voter characteristics that interact with candidate attributes to influence likability judgments and whether those likability judgments have an effect on ultimate vote decisions. We will utilise both a conjoint experiment and a survey experiment to answer our research questions. The conjoint experiment will shed light on which combinations of candidate characteristics are more/less likely to lead to candidates being perceived as likable, while the survey experiment will allow us to consider the relationship between likability judgments, other evaluations and vote choice. We expect to find that likability assessments will vary systematically for candidates with different combinations of gender, race/ethnicity, and partisanship and that this will have direct effects on these candidates' likelihood of receiving subjects' votes.