Liberal and conservative voters exhibit different responses to gender stereotypes of female candidates, according to research conducted by a UC Berkeley political science doctoral student.

Rachel Bernhard — who presented her paper “Do Voters Prefer Well-Behaved Women? Experimental Tests of Competing Stereotypes” at an American Political Science Association conference in San Francisco in September — found that conservative voters were less likely to vote for female and feminine candidates and that liberals were more receptive to female and feminine candidates.

Through an online forum for researchers called Amazon’s Mechanical Turk, Bernhard conducted a survey of 2,666 online respondents, who were all shown a traditionally masculine or feminine biography of a single hypothetical candidate running for an unspecified political office.

Hypothetical masculine candidates were described as working in a moving service and as a Little League coach, while hypothetical female candidates were described as working in a cleaning service and in an adult literacy program.

Bernhard found a link between voters’ ideologies and their likelihood to vote for a masculine or feminine candidate. She found that, on average, conservative voters appeared to prefer political candidates who were masculine, male or both, and that they preferred masculine female candidates more than feminine female candidates because of their embrace of traditional gender roles.

“My hope is that we’ll get people thinking about ways that they might be unconsciously evaluating candidates without realizing it,” Bernhard said. “The study isn’t saying that people are sexist or misogynistic … just that information that we might not be aware is influencing us is influencing us, and that’s relevant to the gender gap that we see in U.S. politics.”

UC Berkeley professor of political science Gabriel Lenz said that Bernhard’s experiment benefited from her carefulness in the way she subtly varied feminine and masculine traits so that respondents — who may not want to appear sexist — were not aware of what they were testing for in the research.

Lenz also said he hopes that Bernhard’s findings will help inform voters of their unconscious biases and inform potential female candidates of how to best appeal to voters.

According to Laura Stoker, UC Berkeley professor of political science, there exists a body of ongoing research that already suggests that voters are not only penalizing female candidates for being women, but also that some sets of voters are responding strongly to information cues related to masculinity and femininity.

Nevertheless, Stoker said the research will open up new avenues for further research into the gender politics and that at this stage, political scientists are not ready to draw real-world inferences.

“This is a new area of research that she’s opening up,” Stoker said. “And she’s beginning to have some interesting findings.”