Kathleen Dolan is a professor of political science at the University of Wisconsin-Milwaukee and the author of the recent book “When Does Gender Matter? Women Candidates and Gender Stereotypes in American Elections.” Her findings contradict a great deal of conventional wisdom about gender and elections. She kindly answered some questions via email. Below is a lightly edited transcript.
How would you describe the conventional wisdom about how gender matters in U.S. elections and for female candidates in particular?
The conventional wisdom is that the presence of a woman candidate disrupts the traditional influences that shape voting behavior. Pundits, the media and campaign professionals tend to think that the sex of a woman candidate is a major consideration for voters and really changes the way women have to campaign. Research by Kelly Dittmar details how much campaign professionals believe that a woman candidate’s sex has to be managed. Some political science and social psychology suggests that people view male and female candidates through the lens of gender stereotypes, which shapes the way they are evaluated and whether people will vote for them.
You think that conventional wisdom is wrong, or at least needs to be qualified.
The conventional wisdom has been shaped by the early research that focused on the presence of gender stereotypes in the minds of voters. However, much of this research is based on experiments involving hypothetical candidates like “Jane Edwards” and “Joe Edwards” or hypothetical “women” and “men” who run for office. Unfortunately, the results of many of these studies have been used as cautionary tales of the challenging environment women candidates face.
However, in the real world, voters almost always know something more about candidates than simply whether they are male or female. These experimental studies don’t tell us anything about whether or not people actually use gender stereotypes to choose among actual candidates.
Tell us about the data you collected for this book.
My data were designed to address this inability to link information about the attitudes people hold, like gender stereotypes, with information about their choices in real elections. I conducted a two-wave panel survey of 3,150 U.S. adults from 29 states during the 2010 election. The sample included respondents who lived in congressional districts and states that had races for the U.S. House, Senate and governor in which a woman was running against a man and other races in which two men were running.
In the first wave of the survey in September 2010, respondents answered questions intended to measure abstract gender stereotypes — whether respondents believed that “women or men who run for office” would be better able to handle particular policy areas and whether they possessed certain personality traits.
In the second wave of the survey, conducted in November, I asked people the same battery of items about policy competence and traits, only I inserted the names of the actual candidates for House, Senate and/or governor that the respondent had experienced. The second wave of the survey also asked about how respondents had voted in various elections.
The goal was to see whether people’s gender stereotypes were related to how they evaluated the specific candidates in their state and district.
What abstract stereotypes do people have about male and female candidates? Are those stereotypes inherently more favorable to male candidates?
My data suggest that gender stereotypes are on the wane. On all eight measures of trait stereotypes and on seven of the nine policy stereotypes, a majority of respondents said there was no difference between male and female candidates in their traits or policy abilities. Only on abortion and child care did a majority of respondents see one sex as better suited to handle these issues than the other, with respondents seeing women as better able than men to handle them.
Among respondents who did see a difference between women and men, their responses corresponded with the expected stereotype. But most people do not hold gender stereotyped attitudes about male and female candidates’ abilities and traits.
Do these stereotypes actually matter at the ballot box?
In a word, no. There were a couple of instances in which abstract stereotypes were related to the evaluations of specific candidates. But there were no circumstances in which gender stereotypes were significantly related to vote choice, both in mixed-sex and in male-only races. Some people may hold gender stereotypes, but they do not employ them when making vote choice decisions.
Instead, the traditional influences on vote choice, namely political party and incumbency, were the most important determinants of vote choice in races with and without female candidates. Respondents in this sample were overwhelmingly likely to vote for the candidate of their party regardless of the sex of the candidate. People vote for women candidates when they share her party and don’t vote for her when they don’t.
You compare people’s abstract stereotypes of male and female candidates to their specific evaluations of the actual male and female candidates that ran in 2010. What’s different when people evaluate the actual candidates?
When people evaluate actual candidates, they are less likely to see no difference between women and men. In terms of policies and personality traits, a majority of respondents could identify a candidate they saw as better than the other. But the candidate seen as “better” didn’t always conform to the expected gender stereotype. For example, people might see the woman candidate of their party as better able to handle foreign affairs, an evaluation that is counter to the traditional gender stereotype of women, but is probably driven by partisan affinity or the record of a particular candidate. In evaluating real candidates, people have more information than just the candidate’s sex and so don’t fall back on gender stereotypes.
How do evaluations of the actual candidates affect vote choice, and does this work to the advantage of either male or female candidates?
Evaluations do influence vote choice: people tend to vote for the candidate they saw as better at some policy issue or more qualified in terms of a particular trait. This was true for both women and men candidates. Positive policy evaluations were a bit more likely to influence support for women candidates than were positive trait evaluations, but both were important. The same general pattern existed for men. But there wasn’t a clear or consistent advantage for either sex. Instead, people tended to evaluate the candidate of their party more positively and were more likely to vote for that candidate.
You also look at whether male and female candidates in 2010 differed in terms of their prior experience, how much money they raise and how well they do on Election Day. Are there important differences here?
There are some differences between the women and men candidates in my sample, but because I only have data from one election cycle, I don’t know if these differences are idiosyncratic or part of an important trend.
Women and men candidates had generally equivalent levels of education, although men running for Senate and governor were more likely to have a graduate degree than were women candidates for these offices. Men who ran for the House were much less likely than the women to have previous elected office experience, while women who ran for the Senate had less prior experience than the men who ran. Campaign spending was the same for House candidates, while women Senate candidates spent more money than the men, and men running for governor spent more than the women. When we compare candidates by seat status — incumbent, open seat, challenger — female incumbents and open seat candidates won at the same rate as did men, and female challengers lost at the same rate as did male challengers.
What about in terms of how the candidates presented themselves? Did male and female candidates campaign in “stereotypical” ways?
In general, they did not. People have often wondered whether candidates contribute to the stereotypes people hold by campaigning in stereotyped ways. But the candidates’ television ads and websites didn’t show big differences based on gender. The bigger differences involved party. Women and men ran on the issues that were most relevant to voters and to their parties in 2010.
In the concluding chapter, you write: “In all, the results of this analysis demonstrate that gender stereotypes are not a significant impediment to the success of women candidates.” Do you think that applies in 2016 as well, especially for Hillary Clinton?
I do. I believe that the findings demonstrate pretty clearly that stereotypes are waning as a way of thinking about candidates. And even when people hold stereotypes, they don’t rely on them. People vote for the candidate of their party and there is no evidence to suggest that Democrats will abandon Hillary Clinton because she is a woman. If stereotypes are a time-saving device for people who have little information about a candidate, we might expect them to be a particularly weak force in 2016, since Clinton is about as well known as a candidate could be.