Identity and Rhetoric: Tracking Emotional Reactions to Presidential Debates

Authors: Kim Fridkin Arizona State University, Sarah Allen Gershon Georgia State University, Jillian Courey Arizona State University, Kristina LaPlant Georgia State University [Under Review]



Delivered at the Annual Meeting of the American Political Science Association, San Francisco, CA, August 31-September 3, 2017


The 2016 presidential contest was historic along a number of dimensions, including the first woman general election candidate and the first general election candidate in history with no political or military experience. The 2016 election season was also distinct in the salience of identity-based discussion, with immigration status, religious affiliation, race and gender all figuring prominently into the candidates’ messages. Furthermore, highly divisive rhetoric and uncivil attacks plagued a contest between two candidates who were disliked by large segments of the electorate. Thus, this unusual election season allows for a unique opportunity to study the impact of candidate messaging during a historic, yet highly polarizing presidential race. We propose that emotion played an important part in shaping reactions to candidate messaging in 2016 and we explore the role of emotion employing a unique study of the presidential debates.

Research tells us that emotion influences how people process messages (Brader, 2005; Isen, 1987; Lodge and Taber, 2013; Marcus, 2000). Yet, given the identity-centered messaging present during the campaign, it is likely that voters reacted differently to candidate’s messages, based on their identity. In particular, we might expect certain racial, gender and ethnic groups to be more upset or enthusiastic in response to particular messaging. For example, female viewers might have more hostile reactions to Trumps’ attack on Clinton as a “nasty woman” in the third debate or racial and ethnic minorities might be more upset by discussion of the stop and frisk policy in the first debate. Given the unusually divisive nature of the presidential debates in 2016, existing research concerning candidate messaging, identity and emotion does not provide a clear guide for understanding responses to Donald Trump and Hillary Clinton or their messages.

In this study we examine how identity shaped emotional responses to the campaign rhetoric by examining responses to the first and third presidential debates of 2016. Through the use of a controlled experiment, we are able to track voters’ change in attitudes and responses to the candidates after exposure to the debates. We further explore emotional responses by using facial expression software to track voter responses to candidates during the debates. Rarely utilized in our field, this facial expression software allows for an unobtrusive measure of emotional responses, recorded in real time during the presidential debates. The software, called “Emotient,” measures the valence of emotions, as well as assessing the presence of distinct emotional states (e.g., Anger, Surprise, Fear, Contempt, Sadness, Disgust).

Leveraging data from both the data generated during the debates by the facial expression software and respondents’ pre/post-test questionnaires, we examine how different groups (e.g., men, women, Whites, Latinos, African Americans,) responded to candidates’ messaging during the debate. We conclude by discussing the implications of our findings.

Preliminary Figures

Screen Shot 2017-08-22 at 7.49.45 PM


Screen Shot 2017-08-22 at 8.25.12 PM

%d bloggers like this:
search previous next tag category expand menu location phone mail time cart zoom edit close