It’s only February 2019, and we already have five mainstream female candidates for president: Democratic Sens. Kirsten Gillibrand (N.Y.), Kamala Harris (Calif.), Amy Klobuchar (Minn.), Elizabeth Warren (Mass.) and Rep. Tulsi Gabbard (Hawaii). Right off the bat, Warren — the first to announce an “exploratory committee” in the race for the Democratic nomination — inspired some textbook examples of how not to think about female politicians. Take Politico’s article “Warren battles the ghosts of Hillary,” published at the end of December, which included the poisonous question: “How does Warren avoid a Clinton redux — written off as too unlikable before her campaign gets off the ground?” It received a drubbing on Twitter and in the press because the two politicians had little in common, aside from their gender; plus, its argument about likability was purely speculative, with no examples of unpleasantness or on-the-record sources proffered.
As more female candidates join the contest, however, the anxieties about the demographics of the race are taking a clearer shape. It’s not about a candidate’s personality anymore. The “problem” involves the electability of women as a class. Yet discussions of whether women face a penalty in elections often rely on a phenomenon you might call “sexism by proxy.” It’s not me — the journalist or party power broker or ordinary citizen — who thinks ambitious women are bossy or unlikable, the argument goes. It’s other voters who (may) hold such views. And in a close election, any disadvantage can make you a loser. So why “take another chance,” as one former Democratic National Committee member told the New York Times, in an article with the headline “ ‘There’s a Real Tension.’ Democrats Puzzle Over Whether a Woman Will Beat Trump.” (Former vice president Joe Biden’s claim that only he is electable against President Trump implies a similar notion.)
Behind all this is the belief that a woman — even an exceptional one — is less likely than a man to win an election against a man. But even if women face a slight penalty at the ballot box (a claim social scientists still argue about), that doesn’t answer the question of what voters should do with that information. If it is unethical to vote against a person purely because of her sex, is it any better to vote that way because other people are sexist? The problem quickly becomes self-fulfilling, especially when commentators amplify the concern. If reporters and analysts make Americans think women are seriously hampered, more voters will write off female candidates. The United States will never get a female president if these are the rules.
People understandably think sexism contributed to Hillary Clinton’s defeat in 2016. Although 92 percent of Americans told told Gallup in 2015 they’d vote for a woman, that still leaves 8 percent who would not. As a candidate, Trump’s sexism was overt, with attacks on Clinton as a “nasty woman,” harsh criticism of female journalists (notably Megyn Kelly), derisive comments about Republican candidate Carly Fiorina and, of course, the “Access Hollywood” video. After the election, an analysis of survey data by Tufts political scientist Brian Schaffner and his team found that sexism motivated some Trump voters.
But it’s not at all clear that sexism made a difference to the election’s outcome. Clinton famously got clobbered by white voters without college degrees, trailing Trump by almost 40 percentage points in that group. A debate immediately sprung up over whether economic worries or racism and sexism drove those votes.
Schaffner’s study found that sexism and racism played far greater roles than economic pressures did — which would seem to lend weight to the sexism-did-Clinton-in thesis. The researchers examined the answers to various questions put to 3,500 voters before the election and 2,830 of those same voters afterward. To gauge levels of “hostile sexism,” the authors asked voters whether women “are too easily offended,” whether women “seek to gain power by getting control over men” and whether women cite discrimination when they lose a fair contest. Similar questions probed racist sentiment, and these voters were also asked how their incomes had changed in recent years.
The most sexist white, non-college voters were 30 percentage points more likely to vote for Trump than the least sexist. (Racism was similarly influential in determining votes.) In contrast, the correlation between economic woes and votes for Trump was much less robust.
But as far as the viability of a female candidate is concerned, the crucial question is whether those voters were ever in play for the Democrats, and Schaffner thinks they probably weren’t. Most voters with hostile sexist views were in the Republican camp, he and his team conclude; they had been moving Republican for many election cycles.
“To have clear evidence that Clinton’s gender cost her the election, we would expect to find that the sexist Obama voters from 2012 switched to Trump in 2016,” he tells me. “That is not what we found in our study.”
In short, polarization may suppress the effects of sexism on voting, he and other scholars think. It’s even possible that Clinton gained from beliefs about in women in politics. Examining some of the same survey data as Schaffner, two University of Texas at Dallas politics professors, Harold Clarke and Marianne Stewart, rated voters on a scale of “progressive” to “traditionalist” on gender issues, analyzing responses to such statements as “women often earn lower salaries than men for doing the same jobs” and “the proper role for women is to be good wives and good mothers.” On a scale of 0 to 100, with 0 meaning most progressive, the median voter scored 40. Traditionalists obviously voted against Clinton, but the general progressive tilt in the electorate on gender issues surprised the authors. Among voters, her identity as a woman “likely helped Clinton more than it hurt her,” they concluded in June 2017.
And a 2018 Pew Research Center study found that while 57 percent of Americans thought women brought different strengths than men to leadership roles (43 percent saw the sexes as roughly similar), overall those respondents saw a slight advantage in the female style. Women are seen as better at working out compromises, for example.
Such findings certainly cut against the conventional wisdom, and are in tension with other studies identifying bias. In one 2008 study, four public-opinion scholars found that 26 percent of voters included “a woman serving as president” on a list of things that made them “angry or upset,” alongside rising gas prices and other irritants. To encourage frankness, the study was designed so the respondents thought the researchers wouldn’t know how they had answered the questions. Granted, that was 11 years and one popular-vote victory by a female presidential candidate ago. But in the run-up to the 2016 election, Fairleigh Dickinson University political scientist Dan Cassino showed that, if you “primed” men to think about gender — specifically, the fact that women increasingly out-earn their husbands — they were more likely to vote for Trump than for Clinton. Male voters who were not prompted to think about that topic preferred Clinton. That study might be cause for worry among Democrats considering, say, a Warren vs. Trump election.
Much of the work on voters’ attitudes toward female candidates understandably focuses on the congressional level and below. (There are 102 women in the House of Representatives — 23 percent of the total — and 25 senators.) And that research tends to suggest that voters are quite open to female candidates. “When women run for office, they raise as much money and are just as likely to win their races as men,” point out Danny Hayes and Alexandra D. Kurtz of George Washington University, and Jennifer L. Lawless of the University of Virginia, in a recent working paper. They also receive similar press coverage, Hayes and Lawless found in their 2016 book, “Women on the Run.” The pair closely analyzed the content of coverage in the biggest newspaper in each congressional district in the midterm elections of 2010 and 2014 and uncovered no evidence that reporters focused on women’s appearance or described their character traits more negatively.
One provocative thesis holds that bias persists, yet women perform as well as men in elections because they are simply better candidates. Sarah A. Fulton, an associate professor of political science at Texas A&M, has described this as the “Running Backwards and in High Heels” phenomenon (an allusion to Ginger Rogers, the backward-dancing partner of Fred Astaire). A 2012 study by Fulton made use of a survey of 2,672 political activists and potential candidates across a random sample of congressional districts in 1998. They were asked about various attributes of incumbents, including dedication to public service, integrity and public speaking. Female incumbents did as well in the 1998 election as men but were rated by observers, on average, far higher than the men on their political skills. If the woman had been merely equally qualified, Fulton wrote, “they would encounter significant electoral sanctions.”
What should the strategically minded voter do with this complex body of evidence? No one is quite coming out and saying that women should not run for the Democratic nomination, though some political discussion and coverage has implied just that. But even if the data was convincing, rather than fragile at best, informally disqualifying women from the nomination would be wrong. Everyone would be horrified if the Democratic National Committee banned women from competing — that would be outlandish, actionable sex discrimination. But is explicitly doing something immoral and illegal really any worse than doing it in a more informal way?
Gillibrand, Harris, and the rest have made the idea of an all-male primary moot. But Democratic and Democratic-leaning voters are still painfully anxious that their party might botch the campaign against Trump. When polled in January, those voters rated two possible Democratic candidates who are white men — Biden and Sanders — the most favorably, citing electability as their main concern.
Electability is a reasonable focus, but such calculations should be informed by data, not instinct. Rather than making female candidates bear the burden of our worries, we should direct our energies elsewhere. First, we should be vigilant that the playing field is indeed level. Despite studies showing that the media is getting better at covering female candidates, voters should remain alert to unfairness. One study of the 2008 Democratic primary race found that major-network reporters referred to Clinton by her first name four times as frequently as they did Barack Obama. In 2016, after a primary debate, commentators on the cable show “Morning Joe” described Clinton as “screaming” and “feisty,” while a New York Times news article said she “appeared tense and even angry at times.” As we’ve seen, when voters are given sexist prompts, they may be spurred to make sexist decisions.
News coverage of Klobuchar in the week of her announcement immediately homed in on on the accusation that she was too hard on her staff, raising the question of whether a male politician would be criticized on similar grounds.
Ultimately, it’s up to individual voters to choose the best candidate, regardless of gender — and not let myths about electability shackle us. The shock over Trump’s victory, in the face of his sexist rhetoric and behavior, makes it understandable that some people think a woman can’t win in 2020. But the best social science research suggests that such a claim is not defensible.
That so many voters persist in underrating the chances of superb, accomplished female candidates, on the basis of sex alone, highlights our lingering disrespect for women. “A woman can’t beat Trump,” it turns out, is the lowest form of political analysis — the prediction you turn into reality by repeating it often enough.