Jonathan Bachman / Reuters

Read enough news reports about Elizabeth Warren’s declaration that she is running for president, and you notice certain common features. In its story on her announcement, The New York Times noted that Warren has “become a favorite target of conservatives” and that, in a recent national poll, “only about 30 percent [of respondents] viewed her favorably, with 37 percent holding an unfavorable view.” The Washington Post observed that Warren’s claim “that she was Native American” has “come under relentless attack from Republican opponents.” It also quoted a Boston Globe editorial that called Warren “a divisive figure.” On CNN, the election analyst Harry Enten suggested that Warren’s “very liberal record, combined with the fact that Donald Trump has already gone after her” has made her a—you guessed it—“divisive figure” whose “favorable ratings are not that high.”

These observations are factually correct. But they also help create a false narrative. Mentioning the right’s attacks on Warren plus her low approval ratings while citing her “very liberal record” and the controversy surrounding her alleged Native American heritage implies a causal relationship between these facts. Warren is a lefty who has made controversial ancestral claims. Ergo, Republicans attack her, and many Americans don’t like her very much.

But that equation is misleading. The better explanation for why Warren attracts disproportionate conservative criticism, and has disproportionately high disapproval ratings, has nothing to do with her progressive economic views or her dalliance with DNA testing. It’s that she’s a woman.

As I’ve noted before, women’s ambition provokes a far more negative reaction than men’s. For a 2010 article in the Personality and Social Psychology Bulletin, two Yale professors, Victoria Brescoll and Tyler Okimoto, showed identical fictional biographies of two state senators—one male and one female—to participants in a study. When they added quotations to the biographies that characterized each as “ambitious” and possessing “a strong will to power,” the male state senator grew more popular. But the female state senator not only lost support among both women and men, but also provoked “moral outrage.”

The past decade of American politics has illustrated Brescoll and Okimoto’s findings again and again. During the 2012 campaign, Republicans attacked Nancy Pelosi in television commercials seven times as frequently as they attacked her Democratic Senate counterpart, Harry Reid. In 2016, the disparity was three to one. Pelosi’s detractors sometimes chalk up her unpopularity to her liberalism and her hometown of San Francisco. But Reid’s successor as the Democratic Senate leader, Charles Schumer, a liberal from Brooklyn, is far less unpopular than Pelosi—and far less targeted by the GOP.

Or compare Hillary Clinton with the men who preceded her as Democratic presidential nominees. In a spring 2016 study, fivethirtyeight.com subtracted the percentage of Americans who felt “strongly unfavorable” toward Democratic nominees from the percentage who felt “strongly favorable” at the same time in the presidential cycle. For almost every nominee from 1980 to 2012, the result was roughly zero. In other words, the percentage of Americans who really liked them equaled the percentage of Americans who really disliked them, which makes sense, given the roughly even nature of America’s partisan split. (The one exception was Michael Dukakis, who, as a little-known governor in the spring of 1988, enjoyed a net positive score of more than 10 points.)

Then, in 2016, everything changed. The percentage of Americans who felt “strongly unfavorable” toward Hillary Clinton exceeded the percentage that felt “strongly favorable” by 20 points. Was Clinton uniquely liberal or uniquely dishonest or uniquely inauthentic enough—compared with the seven male Democratic nominees who preceded her—to explain such a large disparity? Probably not. Gender likely played a key role.

Now the same dynamic is playing out with Elizabeth Warren. Pollsters keep recording her unusually high unfavorability ratings. Last September, CNN found that Joe Biden’s net approval rating among Democrats and Democratic-leaning independents was 30 points. Bernie Sanders’s was 18 points. Warren’s was two points. In December, when Quinnipiac University surveyed Democrats, Republicans, and independents, Biden’s net approval rating was 20 points. Sanders’s was two points. Warren’s was negative seven points. Warren’s relatively low numbers appear driven by her unpopularity among men. When the University of Massachusetts asked state residents whom they support in the 2020 Democratic primary, Warren tied with Biden among women. Among men, she trailed him by 16 points.

These polls often find their way into newspaper articles. The New York Times cited the Quinnipiac survey in its article about Warren’s presidential announcement. On CNN, Harry Enten discussed her polling ratings at length. That’s fine. There’s nothing wrong with journalists discussing public perceptions of a candidate.

The problem is that when journalists ignore what academic research and recent history teach us about gender’s role in shaping those perceptions, they imply—whether they mean to or not—that Warren’s unpopularity can be explained by factors unique to her. They start with the puzzle of her low approval ratings and then, working backward, end up suggesting that her policy views or (pseudo) scandals explain them. Reporters dwell on issues such as Warren’s alleged Native American ancestry not necessarily because they think those issues matter, but because they assume that voters think they matter. If voters didn’t, why would Warren be so unpopular?

What all this ignores is the harsh truth that when women politicians—especially women politicians who embrace a feminist agenda—overtly seek power, many American men, and some American women, react with “moral outrage.” They may not express that outrage in explicitly gendered terms, just as they may not express their anxiety about a black candidate in explicitly racial terms. They may instead cite DNA testing or hidden emails or San Francisco’s cultural liberalism. Or they may simply say they find the candidate’s mannerisms off-putting.

The media’s role is to dig deeper: to interpret these specific discomforts in light of the deeper discomfort that Americans again and again express with ambitious women.

Journalists shouldn’t ignore electability. Elizabeth Warren’s comparatively low approval ratings are a legitimate news story. But the bigger story is that Americans still judge women politicians far more harshly than they judge their male competitors. Unless journalists name that unfairness, they risk perpetuating it.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.