Gender, Voice and Hillary Clinton: The Effect of Candidate Voices

• In Event: The Moderating Role of Gender in Political Communication

Abstract

Articles in the popular press reported that Hillary Clinton’s voice was a big turn-off for many voters during the 2016 presidential campaign. Focus group participants claimed that it was too loud, or too annoying or too much like their nagging ex-wife. Moreover, anecdotal reports suggest that Clinton’s campaign tried to limit the use of her voice in its television advertising. Perhaps there is something distinctive about her voice that makes people recoil. Alternatively, people’s negative reactions to her voice may stem more from their reactions to her as a candidate or their reaction to female politicians generally.

While the effects of Clinton’s voice are difficult to parse out from the effects of Clinton’s appearance, background, and party, it is a worthwhile endeavor given negative perceptions of physiological characteristics may be obstacles for women candidates, generally. Indeed, recent work suggests that perceptions of voice has political implications, with vocal competency (e.g., pitch) exerting an effect on vote choice (Klofstad 2017). Other work suggests that perceptions of a speaker should shift dramatically depending on the medium – whether the message is spoken or read (Schroeder et al. 2017).

We begin with an analysis of the Clinton campaign’s use of her voice in its television advertisements. Drawing on data from the Wesleyan Media Project, we compare how often Clinton’s voice is used to that of the other candidates, and we examine the types of ads in which she is more likely to speak.

To explore the impact of Clinton’s voice, we conduct a 4 x 2 experiment in which we alter, first, the candidate the message is attributed to (Hillary Clinton, a hypothetical female candidate, a hypothetical male candidate and a hypothetical non-gendered candidate) and, second, the method of presentation (a written statement or a verbal statement). Our treatment is an actual statement made by Clinton, which we have both a male and female voice actor read. We also use computer software to create a version of the treatment with an automated voice. We then measure participants’ impressions of each candidate on a variety of characteristics, including overall favorability, positive traits (intelligent, a leader, compassionate, pleasant) and negative traits (annoying, whining).

Our experimental design allows us to measure the impact of Clinton’s voice by comparing the treatment to the written statement by Clinton, thus ensuring that people’s overall impressions of Clinton are controlled for. The experiment also allows us to compare the impact of Clinton’s voice to that of generic male and female candidates and to assess whether the gender of the candidate has more of an impact when read about or heard.

This research is important in that speaks to the role of candidate voice in American politics, a phenomenon that is rarely studied. It provides insight into the outcome of the 2016 presidential race, and it also speaks to the role of candidate gender in American campaigns more generally.

Authors

• Travis N. Ridout, Washington State University

• Kathleen Searles, Louisiana State University

• Patricia Strach, University at Albany, SUNY

• Katherine Zuber

• Erika Franklin Fowler, Wesleyan University