暗网吃瓜

A Rochester undergraduate discovers that prospective voters perceive female politicians as more extreme than their male counterparts.

Morgan Gillespie 鈥�23 had more than a hunch.

Combing through literature on gender bias among American voters, Gillespie, a political science major at the 暗网吃瓜, found a curious gap. While many researchers agreed that a candidate鈥檚 gender influences prospective voters鈥攖he nagging question of how and in which ways had not been answered satisfactorily. Moreover, she found little scholarship that sought to measure voters鈥� perceptions of ideological differences between men and women seeking public office.

That鈥檚 why, with some help from her honors thesis advisor , an associate professor in the , she devised and conducted an experiment that examines the connection between a politician鈥檚 gender and voters鈥� perception of that politician鈥檚 policy preferences.

Gillespie鈥檚 findings are preliminary. This summer, she and Abramson will widen the study and prepare a paper that they hope to publish in a top peer-reviewed journal. The data from her experiment will serve as the foundation. As her advisor聽puts it, Gillespie has produced聽鈥渁 kick-ass honors thesis鈥澛爐hat聽will serve as聽鈥渁聽pilot聽for a more comprehensive experiment based on a nationally representative sample.鈥�

What Gillespie has found so far is striking.

Her results show that voters (both male and female) use gender cues to form beliefs about which policies a politician supports. When voters are unaware of a politician鈥檚 party affiliation, women politicians are seen as more liberal than otherwise identical men. However, when party affiliations are known, female candidates are seen as more ideologically extreme than male candidates of the same party. Furthermore, what has widely been labeled 鈥渨omen’s issues,鈥� or issues that women candidates are more likely to support鈥攕uch as abortion, paid family leave, or the gender wage gap鈥攁re perceived as more ideologically liberal issues or extreme positions. Gillespie found very little difference between male and female respondents in her experiment.

If the findings bear out, Gillespie will have provided evidence against the conclusions of an from 2020, according to which voters appear to prefer women candidates for public office.

The studies on which the meta-analysis is based 鈥渁re conflating a preference for women with beliefs about the candidate鈥檚 ideology or party affiliation,鈥� Gillespie says. 鈥淭he conclusions from these previous studies misinterpret the positive effect of gender.鈥� She notes carefully that her and Abramson鈥檚 findings suggest an additional source of bias that these experiments聽鈥渇ail to uncover.鈥�

Departmental program identifies promising undergraduate researchers

Gillespie, who came to Rochester from San Diego, California, is part of the highly selective that enables senior undergraduates to conduct original social science research in a small, collaborative setting. Stellar students are invited into the program by the department faculty, generally in the spring of their junior year. As Abramson tells it, Gillespie was a shoo-in.

鈥淪he鈥檚 clearly one of the best undergraduates in our department,鈥� he says, pointing to her abilities as a researcher, coupled with her strong work ethic and insightful questions. 鈥淭he progress she makes week to week is more than I expect from most graduate students.鈥�

Gillespie has broad intellectual interests that made her pursue minors in international relations, legal studies, and art history. She鈥檚 also deeply involved in numerous campus activities, including the student government鈥檚 judicial council, mock trial, the pre-law society, the Gamma Phi Beta Sorority, and the women鈥檚 rugby team. As if this were not enough, she has been working at the New York State Office of the Attorney General where she works in mediation for the .

Measuring for gender bias

In order to quantify bias based on a candidate鈥檚 gender,聽Gillespie聽created six hypothetical candidate vignettes such as the one below, into which she randomly assigned pronouns聽(she or he) to provide gender cues.

For a portion of these fictitious candidates, Gillespie randomly assigned a specific party affiliation鈥攅ither Republican or Democrat鈥攖o measure the effect of partisanship. Then she asked a sample of 506 American respondents to predict the candidates鈥� vote choice鈥攜ay or nay鈥攁cross a list of 18 electoral bills.

Gillespie asked the respondents where they thought the six candidates stood on a wide range of issues, such as climate change, police accountability, the safety of police officers, immigration, free school lunches, abortion, the health of incarcerated women, gun safety, minimum wage, national health insurance, support for the armed forces, national security, the gender wage gap, paid family leave, voting restrictions, and access to higher education.

With the help of her advisor, whose research aligns with her topic, Gillespie estimated parameters of a 鈥攕imilar to those models commonly used in the analysis of legislative roll-calls鈥攚hich typically explain a choice between two or more distinct options that are mutually exclusive.

The graphic shows respondents鈥� estimates for candidate positions without having been provided a candidate鈥檚 party affiliation. Here the fictitious female candidates (pink line) score as 鈥渕ore liberal鈥� (to the left of zero) than their fictitious male counterparts (blue line). (Source: Morgan Gillespie)

The graphic shows respondents鈥� estimates for candidate positions after having been provided with a candidate鈥檚 party affiliation. The gap between the perceived positions of fictitious female and male candidates now decreases, yet in both instances female candidates are seen as more extreme: Democratic women (pink line D) are located politically to the left of their Democratic male counterparts (blue line); while Republican women (pink line R) are located politically to the right of their male counterparts (blue line). (Source: Morgan Gillespie)

The results boil down to a double-whammy effect for women candidates at the ballot box, regardless of their party affiliation. First, in a political culture in which centrism is associated with greater electability, female politicians are seen as more extreme than their male counterparts. Second, the issues that women may support are also viewed as more ideologically liberal. Yet, female candidates may face strong pressure to campaign on issues that are more broadly supported by women, or that are categorized as 鈥渨omen鈥檚 issues.鈥� Taken together, Gillespie says, these two biases can make female candidates appear more extreme and thus less electable than otherwise identical male candidates.

The road ahead

The project has proved potentially life-altering for Gillespie. Originally, she had planned on applying to law school after graduating from Rochester this May. But once she got started on the research for her honors thesis, things began to shift.

鈥淚 realized research was something that I had the skills for, something that I learned here at Rochester,鈥� says the 21-year-old. That skill building didn鈥檛 happen in the classroom alone but also came with leadership roles across campus organizations and clubs, which built her confidence in 鈥渃ommunicating my ideas and advocating for myself.鈥�

Upon graduation Gillespie will head to the East Coast for a year as a political science research associate at the Massachusetts Institute of Technology (MIT), with the ultimate goal of pursuing a PhD in political science.

鈥淚 realized that if I applied and pushed myself, I would be able to investigate issues that I always had questions about but that I didn鈥檛 think I鈥檇 be able to find the answers for,鈥� says Gillespie, summing up her experience with her honors thesis research. With that came the realization that 鈥減eople will stand behind me and support my ideas.鈥�

Read more