AMES, Iowa – Gender stereotypes and biases still influence voters, especially in elections with more than one woman on the ballot, according to new research from Iowa State University.
Tessa Ditonto (lead author) and Dave Andersen, both assistant professors of political science, say gender had the greatest effect on down-ballot races, in which women were running for a legislative office and another woman appeared on the ballot for a higher office, such as governor or president. The work is published in the Journal of Women, Politics & Policy.
Ditonto and Andersen conducted two separate studies using computer-simulation software, which allowed them to mimic a real election environment. They presented study participants with candidate information and measured how they evaluated and voted for women candidates. For one study, participants saw as few as zero and as many as eight women on the ballot. Researchers manipulated the candidates for each simulation so that all characteristics were the same, except gender.
“When there was only one woman, gender didn’t matter. People liked her the same as the male candidate and were just as likely to vote for her,” Ditonto said. “When we added another woman, especially one running for higher office, the woman lower on the ballot had more negative evaluations and was less likely to receive votes.”
The researchers found in one scenario gender even rivaled party affiliation. In a simulated race for U.S. House, the likelihood of in-party supporters voting for a woman candidate – over a man running for the other party – dropped from 98 to 74 percent when more women appeared on the ballot.
With a record number of women running in this fall’s midterm elections, Ditonto and Andersen say their findings are concerning, but not discouraging. The research shows women are just as likely as men to be elected, but the context of the election cycle makes a difference. With that in mind, women must recognize gender is still an issue they will confront on the campaign trail, Andersen said.
“Campaigns for lower-ballot candidates can’t dismiss the pervasiveness of gender,” he said. “They have to get their candidate ready for questions about whether she’s tough enough, if she is capable and competent, and understand those questions are not going to go away.”
However, with time those questions may become less relevant. Ditonto says the stereotypes are salient, because women are not the “norm” in politics and there is an expectation that our leaders are men. She sees this gender effect as a temporary bump in the road.
“Hopefully, as more women run for office, gain exposure and win, this effect will go away and won’t be a permanent fixture of our political system,” Ditonto said. “It should be more of a growing pain as more women enter the political world.”
In past studies, Andersen and Ditonto found voters were less likely to seek information about down-ballot candidates and spent more time learning about candidates for higher office. When voters looked for information on women candidates, it often related to their competency and intelligence. To better understand gender effects in elections, more experimental studies, similar to the ones described above, are needed, Ditonto and Andersen said.
In a separate paper recently published in the journal Political Analysis, they found computer simulations with high levels of information more effectively replicated election conditions to test voter behavior. They say the key takeaway from the paper – information matters. When voters were given relevant information about a candidate, they relied more on information and less on gender cues.
“If voters only know a candidate’s gender, stereotypes dominate and they may guess at where the candidate stands on certain issues,” Andersen said. “The more information voters learn about a candidate, the less gender matters.”