Since Barack Obama emerged as a serious challenger to Hillary Clinton for the Democratic presidential nomination, the primaries have become, in part, a referendum on whether Americans are more prepared for a woman or a black man in the White House. The voting has been parsed for signs that the candidates are drawing supporters beyond their particular "minority" " demographic. Over the past month and a half, the feminist pioneers Gloria Steinem and Robin Morgan have both published widely talked-about essays arguing that Clinton would have long since sewn up the nomination if not for the stubbornness of our national sexism. And when Clinton's primary victory in New Hampshire last month caught everyone by surprise, some analysts suggested that the polls had been so wrong beforehand in part because voters in the overwhelmingly white state had been reluctant to share their true, race-based reservations about Obama.
The discussion so far has been rather short on data. There have been surveys asking whether Americans would vote for a black or female candidate for President -- according to a December 2007 Gallup poll, 93 percent and 86 percent, respectively, say they would. Those answers should be interpreted with some skepticism, however, because people are often unaware of their biases and don't tend to reveal them honestly in surveys.
But turn away from the campaign trail, and toward the laboratories where psychologists work, and a fascinating portrait of the primaries emerges. For decades, researchers have been probing bias -- how it arises, how it changes, how it fades away. Their work suggests that bias plays a more powerful role in shaping opinions than most people are aware of. And they suggest that the American mind treats race and gender quite differently. Race can evoke more visceral, negative associations, the studies show, but attitudes toward women are more inflexible and -- to judge by the current dynamics of the presidential race -- ultimately more limiting.
"Gender stereotypes trump race stereotypes in every social science test," says Alice Eagly, a psychology professor at Northwestern University.
It would be a gross oversimplification to reduce the Democratic race to the white woman versus the black man. Factors like Obama's eloquence and inexperience and Clinton's policy mastery and her association with the ambivalent legacy of her husband have played a larger role in how the race has been talked about. And indeed, this presidential contest can be seen as the country's attempt to lurch beyond a blinkered, monolithic identity politics.
But in a campaign in which it's hard to find many substantive policy differences between the leading Democratic contenders, it's notable how well the psychological research on bias predicts the race we've seen so far. Obama's ability to disarm the initial reservations of an increasing number of white voters as the race has progressed -- especially over the past week, in his string of eight straight primary victories -- fits with the findings of bias researchers that racial bias is strikingly mutable, and can be mitigated and even erased by everything from clothing and speech cadence to setting and skin tone.
As Clinton has discovered, gender stereotypes are stickier. Women can be seen as ambitious and capable, or they can be seen as likable, a host of studies have shown, but it's very hard for them to be seen as both -- hence the intense scrutiny and much-debated impact of Clinton's moment of emotional vulnerability in a New Hampshire diner last month.
As the race moves toward the possibly decisive March 4 primaries in Ohio and Texas, Clinton and Obama will have to continue to negotiate the complex demands of campaigning for an office that has been held by an unbroken string of 43 white men. But while this presidential campaign has proven a stage on which these issues can dramatically play out, they also run deeply through the rest of our society. And if the ample literature on bias shows anything, it is that, for all the difficulties Americans have with race, it may prove that attitudes about women are the hardest to change.
. . .
Race and gender are both traits that we cannot help but notice. One hundred milliseconds after we have first laid eyes on someone, we have made a determination about their race; 50 milliseconds later, we have determined their gender. But the reactions are not identical.
When psychologists talk about bias, they use three technical categories: stereotyping, prejudice, and discrimination. Stereotyping is cognitive bias, the tendency to ascribe people a set of traits based on the group they belong to (e.g., "black people are good at sports," "Jews are cheap"). Prejudice is an emotional bias, disliking someone because of their group identity. And discrimination is how we act on the first two.
Sexual prejudice isn't terribly common -- male chauvinists don't dislike women, they just have particular ideas about their capabilities and how they should behave -- but with race, stereotypes tend to go hand-in-hand with prejudice.
Many studies have shown the prevalence of negative associations among white Americans toward blacks. Mahzarin Banaji of Harvard and Anthony Greenwald of the University of Washington have done influential work showing that most whites, whatever their professed racial attitudes, are quicker to associate positive words with images of whites, and quicker to associate negative words with blacks. The test they developed, the Implicit Association Test, or IAT, has become one of the most common tools for measuring bias.
Joshua Correll, a psychology professor at the University of Chicago, measures bias in a more dynamic way, looking at associations with danger. In one set of studies he had mostly white participants play a primitive video game in which they had to make split-second "shoot/no-shoot" decisions based on whether the figure on the screen was holding a gun. Most subjects, he found, were more trigger-happy when presented with an image of a black man.
But follow-up studies have also shown that these biases can be sharply reduced, and in some cases even erased. When participants, for example, are shown images of well-liked black public figures before taking the IAT, their anti-black biases disappear.
"We're finding that racial stereotyping and prejudice are extremely contextual," says Correll. "You can see real reductions in prejudice, and sometimes it actually reverses," crossing over into a sort of stereotypic affinity.
And this, Correll argues, works to the advantage of someone like Obama. "You look at Obama, and he represents himself incredibly well," Correll says. "There are a whole lot of contextual cues that tell us this is someone you don't need to worry about."
The pollster John Zogby sees some signs that white voters have grown more comfortable with black candidates. He offers the example of Harold Ford, the young, black Democratic congressman who narrowly lost his bid for one of Tennessee's US Senate seats in 2006. Traditionally, Zogby points out, black candidates do worse on Election Day than in pre-election polling because people tell pollsters they're more comfortable with black candidates than they actually are -- this phenomenon, the so-called Bradley Effect, is what some analysts thought helped Clinton last month in New Hampshire. But, Zogby points out, Ford actually did better in the final vote than in pre-election polling, suggesting a dissipation of the Bradley Effect.
Some of the most dramatic work in racial bias mitigation was published in 2001 by John Tooby and Leda Cosmides, a husband-and-wife team of evolutionary psychologists at the University of California, Santa Barbara, and their then-student Robert Kurzban. In their study, they presented participants with a series of images of people, each with a sentence that the person in the image had supposedly said. Later on, the test subject would be asked to recall who had said what.
What they were after were wrong answers. The ways in which test subjects misattributed quotes betrayed the categories by which they grouped people. Subjects, for example, were far more apt to misattribute something one black man had said to another black man, rather than to a white man or to a woman.
Surprisingly, though, the researchers found that they were able to get people to stop paying attention to race with a simple manipulation: they showed images of people wearing one of two colors of T-shirts, paired with quotes that gave the impression that the T-shirts correlated with membership on different "teams." In response, test-takers started grouping people on the basis of the T-shirt color rather than their skin color, confusing T-shirt "team members" of different ethnicities with each other.
And while the study wasn't looking at bias, the implications are clear. "If you're going to discriminate on the basis of race you first have to notice it," says Kurzban, now at the University of Pennsylvania. In an experimental setting, at least, he argues, you can get people to stop doing that.
The researchers didn't see a similar effect for gender. According to Tooby, "People can cease to notice ethnicity as a factor in how they conceptualize somebody in a way that they don't seem to be able to with gender."
. . .
There is work suggesting that implicit gender stereotypes can also display a degree of mutability, at least among women. Studies conducted by Nilanjana Dasgupta, a psychology professor at the University of Massachusetts, Amherst, have found that exposing women to photos and biographical information about accomplished women like Meg Whitman, CEO of
Still, psychologists specializing in gender bias say that many studies have shown how strong a force gender stereotyping is.
In one particularly telling strain of research, called the Goldberg paradigm, two sets of participants are asked to comment on something, perhaps a resume or a speech or a work scenario in which a boss speaks with an employee. To one audience, the person involved is described as a woman, in the other he is a man. Time and again, male participants (and, in some cases, women as well) judge the resume more harshly if it is a woman's, or say the speech was strident if given by a woman but assertive if given by a man, or that the female boss was pushy while the male boss was concerned.
Women in these studies are typically judged to be less capable than men with identical qualifications, but it's not impossible for them to be seen as competent. The problem is that if they're understood to be capable, the majority of respondents also see them as less likable.
"The deal is that women generally fall into two alternatives: they are either seen as nice but stupid or smart but mean," says Susan Fiske, a psychology professor at Princeton who specializes in stereotyping.
And unlike racial bias, there's little evidence that these attitudes are softening.
According to Eagly of Northwestern, the problem isn't that women aren't traditionally understood as smart, but that they traditionally aren't understood to be "assertive, competitive, take-charge" types. More than intelligence, she argues, this "agentic" quality is what we look for in leaders, and, as both surveys and experimental studies have shown, we find it deeply discomfiting in women.
"That's what Hillary Clinton is up against," argues Eagly. "She's had to show her toughness, then people turn around and say she's too cold."
Amy Cuddy, a psychologist at Northwestern, suggests that the durability of gender stereotypes stems in part from the fact that most people have far more exposure to people of the opposite gender than to people of different races. As a result, they feel more entitled to their attitudes about gender.
"Contact hasn't undermined these stereotypes, and it might even strengthen them," she says. "Many people don't believe seeing women as kind or soft is a stereotype. They're not even going to question it, because they think it's a good thing."
Tooby takes a more biological view. As he argues, in the prehistoric environment in which our brains evolved, race had no meaning -- no one could travel far enough to meet anyone who didn't look like them. Gender, on the other hand, meant a lot. It predicted what someone's status would be, what their priorities were, whether they were a potential rival or a potential partner.
Indeed, the only other trait that we notice as strongly as gender, Tooby points out, is age. Clinton is 60 years old, Obama 46. And no matter who wins the Democratic nomination, the face-off against the 71-year-old John McCain may introduce a whole new aspect to the identity politics of the campaign.
Drake Bennett is the staff writer for Ideas. E-mail firstname.lastname@example.org.
Correction: Because of a reporting error, a story in Sunday's Ideas section about bias and the presidential election incorrectly stated the number of men who have been president. The correct number is 42.