Upcoming Events

"Does the Medium Matter? Cueing Identity in Candidate Experiments"

Bio:

Sara Morell’s research is at the intersection of gender and American politics. Her current book project considers the relationship between women candidate training organizations and increases in women’s political candidacies in the United States. This project demonstrates that the gendered signal sent by women’s candidate training organizations increases women’s beliefs that these organizations will address their material and psychological barriers to running. This has implications both for women’s individual-level ambition and state-level representation. Her research has been published in the British Journal of Political Science and she has received grants from Time-Sharing Experiments for the Social Sciences (TESS) and the American Political Science Association’s Women, Gender, and Politics Research Section. At The College of New Jersey Sara teaches courses in American Government, Political Analysis (a course on the foundations of political science research), and Gender Politics.

 Abstract:

How do different experimental designs that describe political candidates influence how respondents process information about candidate’s identities? Traditionally, studies looking at how a candidate’s race, ethnicity, and gender influence vote choice have used vignette experiments, where respondents receive a paragraph candidate biography and researchers vary relevant characteristics (race, gender, age, traits, etc.). However, there has been a growing shift towards using conjoint experiments to study voter preferences toward political candidates. In conjoint experiments, information is presented in a table, allowing researchers to evaluate decision-making in a multi-dimensional context. We argue that the shift from presenting candidate information in a vignette paragraph to a conjoint table also influences how characteristics are understood, processed, and used in candidate evaluations by survey respondents. Our first experiment evaluates whether differences in the treatment mode (vignette paragraph or conjoint table) used to present candidates’ race, ethnicity, and gender, impact how respondents evaluate political candidates. Our second experiment then tests the potential mechanisms for this variation, considering whether stereotype control, information salience, and cognitive load explain variation in candidate vote choice across treatment mode. This research critically informs our understanding of the role of research design decisions in estimating the impact of candidate identity on voter preference.