Policy

Think you're not racist?

Research uncovers our secret prejudices, and ways to overcome them

By Alice G. Walton     
June 17, 2014

From: Magazine

Photo by Josh Stunkel.

A poorly worded Twitter message landed the US Republican National Committee in trouble in December 2013. “Today we remember Rosa Parks’ bold stand and her role in ending racism,” the RNC tweeted. The predictable backlash was rapid and furious, noting that racism was, in fact, far from over. Hours later came a correction: “Previous tweet should have read, ‘Today we remember Rosa Parks’ bold stand and her role in fighting to end racism,’” the official handle noted.

The RNC may have stumbled, but there are some indications that racial discrimination has lessened. In recent decades, African Americans have filled such high-profile positions as US attorney general, president of an Ivy League university, CEO of a Fortune 75 company, US secretary of state, and, of course, US president. Many observers see these developments as a sign of a vastly improved racial climate in the United States. 

It has been 50 years since the Civil Rights Act outlawed discrimination based on race, color, religion, sex, or national origin. The landmark legislation marked the end of the era of legalized racism. Now some affirmative action programs, created to encourage and promote diversity and the presence of underrepresented minorities, are being rolled back.  

However, while overt racism may be on the wane in the US, research suggests it remains just below the surface. Very few people would admit to being biased, yet there’s strong evidence that biases continue, often under the level of our expression and of our awareness. By analyzing available data, and running controlled experiments in labs, scientists are spotlighting where racism remains. Perhaps most importantly, they are offering some ideas of how to eradicate it. 

chartHow white is your resume?

Ten years ago Marianne Betrand, Chris P. Dialynas Distinguished Service Professor of Economics at Chicago Booth, and Sendhil Mullainathan, then at MIT, published a famous study entitled, "Are Emily and Greg More Employable Than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination," in which 5,000 fictitious resumes were sent in response to 1,300 job postings in Chicago and Boston. The resumes were either “high quality” or “low quality,” varying in the typical things that set resumes apart—job and internship experiences, academic institutions, and languages spoken. Then, the team randomly assigned either a “white-sounding” name, such as Emily Walsh, or an “African American–sounding” name, such as Lakisha Washington, to each resume.  

The results were unambiguous. White-sounding applicants got 50% more callbacks than African American–sounding candidates. This didn’t seem to be a matter of how common the names were or the apparent social status of the applicant, but simply a function of what the names suggested about the race of the fictional applicants. Even more disturbingly, white applicants with higher-quality resumes had a strong advantage over their African American peers. The authors suggest that this makes it less enticing for African Americans to develop high-quality resumes, which makes hiring discrimination part of a destructive cycle. 

In a 2012 paper, “Do Judges Vary in Their Treatment of Race?” Bertrand finds a similar “racial gap” in jail sentencing. Bertrand, Mullainathan (who’s now at Harvard), and the University of Pennsylvania’s David Abrams took advantage of the fact that cases are randomly assigned to different judges in the criminal-justice system, to determine whether judges vary in sentencing African Americans and whites. The researchers find a clear variation between judges in the size of the “race gap”: some judges are more likely to sentence African Americans to longer terms. “We find evidence of significant interjudge disparity in the racial gap in incarceration rates,” the authors write, “which provides support for the model in which at least some judges treat defendants differently on the basis of their race. The magnitude of this effect is substantial.” 

The problem is that the study can’t tell which way the trend goes—that is, whether discrimination or reverse discrimination might be at play for any given judge, as the researchers don’t know the race of the judges. But whichever the direction, the larger issue is that we don’t know on what level the discrimination is happening. Just as with the human-resources personnel in Bertrand’s earlier study, it’s hard to know whether racial discrimination is happening explicitly (consciously) or implicitly (subconsciously). 

However, there’s some good evidence that forms of discrimination including racial bias occur outside of our awareness, and Bertrand has done extensive work in what’s known as implicit discrimination. “It may be that the recruiters don’t read any further than the name,” she says, referring to her labor-market-discrimination study. “We really don’t know what level it’s happening on. But it doesn’t have to be a conscious choice.” In other words, if a human-resources professional has a pile of resumes to get through, inferring information from the name might be a timesaving maneuver, so that even if the name is just one part of the, “Is this candidate worth a callback?” question, it could be computed quickly and subconsciously. 

Judges deciding on sentencing may have longer to make their decisions, but they likely rely on implicit processing as well, since they, too, must be aware on a conscious level that they have to be race-blind. “No judge is likely to acknowledge, on his or her own, ‘Well, of course I take race into account,’” says Abrams. “And likely they don’t in any explicit way. But they probably do implicitly, because we take all kinds of things into account implicitly. And I think making judges aware of it [their implicit racial bias] could potentially help going forward.”

The suggestive power of counterstereotypes

Other researchers are starting to learn more about the implicit and explicit levels on which these processes occur. Jane L. Risen, associate professor of behavioral science at Chicago Booth, finds some strong evidence that though we may deny it, we're highly susceptible to what's known as incidental exposure, even when it comes to race. Put another way, we’re very likely to pick up on messages to which we’re exposed only very briefly, and which we may be barely aware of, if at all. 

Risen’s research, done in collaboration with the University of California, Berkeley’s Clayton Critcher, was in large part sparked by the election of Barack Obama to his first term as president. She explains, “Obama’s election was an inspiration for many, truly a symbol of hope. But, at the same time, there was a narrative suggesting that his election somehow proved that we live in a postracial world where African Americans face no disadvantages.” Risen is referring to CNN commentator William Bennett’s remark on the apparent significance of Obama’s election into office: “I’ll tell you one thing it means, as a former secretary of education,” he said, “You don’t take any excuses from anybody who says the deck is stacked.” 

Of course, one person’s success does not tell us much about the state of race relations. In fact, the opposite argument could be made. Risen and Critcher write, “. . . the fact that it is so notable for a black man to be a mainstream candidate for president could itself symbolize how race continues to be an important and potentially limiting factor in modern society.” In the same way that George Burns’s cigar-smoking habit was not a testament to the harmlessness of cigars over the course of a very long lifetime, but rather a remarkable exception, so too the election of a mixed-race candidate to the presidency may be an exception rather than a testament to a postracial world. Obama may have been elected not because the deck wasn’t stacked, but in spite of it being stacked. 

To examine the inferences that people make when exposed to examples of successful African Americans, such as Obama, Risen and Critcher developed a study in which participants thought they were involved in two unrelated experiments. Unbeknownst to them, the first part of the experiment served simply to expose them, in the form of a celebrity-trivia quiz, to pictures of high-profile, successful individuals. For some participants, one or two of the individuals included in the quiz were African Americans in “counterstereotypical” positions, such as politics and academia. (Other participants only saw pictures of white people.) “We showed them two pictures of African-American people, and asked, for example, ‘Which is Democratic hopeful Barack Obama?’ But both were pictures of black people. We didn’t care what their answers were.” The researchers just wanted the participants to be reminded that the high-profile people depicted were black, to give them incidental exposure.  

Then the team asked the participants, in a number of different ways, whether race was currently a limiting factor for people in the US. They found that people who were incidentally exposed to counterstereotypical African Americans, such as Obama, former Brown University President Ruth Simmons, or Nobel laureate Toni Morrison, were much more likely to say that race wasn’t a limiting factor. “Those exposed to Black counterstereotypical exemplars were more likely to deny racism and state that Blacks could pull themselves up by their bootstraps if they tried,” the team writes. 

Interestingly, it was solely the counterstereotypical nature of the exemplars that was key to the effect—highly successful African Americans in more “stereotypical” professions didn’t provoke it. When participants were shown pictures of singer Diana Ross or basketball player LeBron James, they did not later issue the same denial. 

This all happened implicitly—the participants didn’t realize that looking at the pictures affected their opinions. In fact, when explicitly asked, participants said these counterstereotypical examples didn’t indicate a positive racial climate. They “stated that the exemplars did not reflect the decline of racism, did not even remind participants of racism’s decline, and certainly should not be taken as evidence that race was not an impediment to success,” says Risen.  

So, participants display consistent and commendable logic—explicitly, anyway. But implicitly, it’s another story. Over and over again, incidental exposure to the counterstereotypical examples affected people’s opinions, prompting them to say race was no longer a disadvantage; it even prompted them to indicate that policies designed to address racial inequality were no longer necessary. In fact, Risen says, even when the team specifically told the participants that exemplars were exceptions to the rule, the participants’ opinions about race were still swayed. “The effects even emerge when you specifically note that this is the only person who’s accomplished such a thing. For example, even when you tell participants, ‘Kenneth Frazier is the ONLY Fortune 75 CEO that’s black,’ you still get the effect.” 

The team concludes that there is “no evidence that automatic inferences unfolded with intention or awareness,” which implies there’s a big disconnect between explicit and implicit processes. 

chartRational people are more sensitive

The experiment revealed another significant finding: people who most vehemently argue—explicitly—that a single successful example does not say much about the status of race relations, are the very ones most swayed by incidental exposure.

Risen and her team first gave participants tests to measure their level of thoughtfulness, using the “need for cognition” (NFC) scale, which tests for how thoughtfully and rationally one approaches the world. Then the researchers repeated the earlier experiment. It turns out that people who ranked high on the NFC scale—those who approached the world in a more thoughtful way—were much more likely to be affected by the African American primes, or incidental exposures, and to draw the conclusion that race was not a limiting factor. “In fact,” says Risen, “we found that those participants who most vehemently objected to the explicit logic, those who claimed to be most interested in careful, rational thinking, were the ones most likely to show an automatic effect of exposure.” 

Risen says that the results illustrate the mind’s remarkable ability to keep separate explicit and implicit cognitive processes. Some people, the highly thoughtful ones, are particularly sensitive to incidental exposure: they don’t just see an isolated example of an African American person in a high-profile position—they see the example as commentary on the way the world is. Says Risen, “these highly thoughtful, high NFC people, who are less affected by blatant primes, are more influenced by subtle primes. They’re paying attention to their environments . . . They’re making inferences about subtle primes that might get missed by other people.” 

To test whether the tendency to approach the world with an inferential mindset is what makes people especially susceptible to the influence of exposure, the researchers also manipulated people’s mindsets. Previous research has shown that people are receptive to the patterns of thinking they’re trained in—for example, when they’re trained to think abstractly or concretely in an initial test, they’ll take that with them, and approach subsequent problems the same way. So, Risen and Critcher had a group of participants answer questions from the Law School Admissions Test, or LSAT, that are designed to strengthen “inferential thinking.” Then the participants repeated the earlier parts of the experiment, and they were much more likely to score higher on the denial of racism test than those who were not “trained” in inferential thinking. 

Just like the more thoughtful people were more likely to pick up on the incidental cue and therefore deny race was a limiting factor, people who were trained to place more emphasis on reasoned thinking were similarly likely to deny racism. That is, people were more likely to make inferences about race relations after incidental exposure when they were in the habit of making inferences. 

Taking subtle cues from the environment is both a blessing and a curse, at least in the case of race: the downside is that we are apt to take the “African American president” cue as a sign that race is no longer a limiting factor. It seems to be seen as a marker of a goal achieved, not a work in progress. Therefore, the drawback of implicit learning is that we see these notable occasions—Obama’s election or Morrison’s Nobel Prize—as symbols that we’ve solved the problem, rather than milestones in an ongoing process. 

However, incidental exposure has huge potential to help shift beliefs in more productive ways, too. “Being exposed to a black president isn’t a bad thing,” says Risen. “The answer is not, ‘Don’t expose people to these exemplars.’ Although it can lead to the problematic inference that race relations are already solved, it may also lead people to make positive inferences about members of the group.”

In other words, we may be able to use the brain’s robust capacity to pick up on subtle cues. The answer of how to best do this isn’t entirely clear, but keeping the benefits and potential pitfalls in mind as we create public-policy initiatives, develop televisions shows and movies, and write advertising campaigns is a place to start. 

chartWould you vote for the darker-skinned candidate?

While we make conscious choices regarding our personal and political beliefs, unconscious cues shape the values that frame these beliefs—and can affect our opinions about race.

In the 2008 US presidential election, Democratic presidential hopeful Hillary Clinton's team was accused of darkening the skin tone of her opponent Barack Obama in a television ad, hoping to trigger unconscious cues. In 1994, Time magazine came under fire for darkening a picture of American football player O.J. Simpson after his arrest. But Eugene M. Caruso, associate professor of behavioral science at Chicago Booth, suggests that viewers themselves may subjectively "darken" a person's skin tone, depending on whether their political beliefs match those of the individual. 

In a study designed with Nicole Mead, now at Rotterdam School of Management, and New York University’s Emily Balcetis, Caruso asked participants to read about a hypothetical political candidate whose platform either agreed or disagreed with the participants’ own beliefs. The candidate was also biracial, so there was some ambiguity about his skin tone. Participants were shown three different photos of the person—one had been lightened, one darkened, and one left unchanged. The participants were asked to choose the picture that best represented the person they’d just read about.  

Participants—almost all of them white—generally chose the lightened picture of the candidate if their political beliefs aligned with the candidate’s. If their beliefs were not aligned, they tended to choose the darkened picture. 

In subsequent phases, Caruso and his team used darkened, lightened, and untouched pictures of Obama, and asked participants which one was most representative of him. More-liberal participants chose the lightened picture, while those who were more conservative chose the darkened photos. Regardless of their party affiliation, people who said they intended to vote for Obama were more likely to choose the lightened pictures of him compared to those who said they did not. 

A person’s political beliefs may impact how supportive he is of, say, Obama’s health-care plan. That’s understandable, says Caruso. What’s more surprising is that a person’s political beliefs could affect how he or she perceives skin tone. Caruso compares this to having a liberal and conservative look at the same physical copy of a health-care plan sitting on a desk before them, but disagree over its thickness. “Even something we feel we should see similarly, like a person’s racial identity or physical characteristics, can be influenced by our desire to see that person favorably or unfavorably,” he says.   

But where does the dark = bad, light = good divide come from? One theory offered by the late cultural anthropologist Margaret Mead is that it’s an evolutionary thing: dark skin makes people think of night, which can be an uncertain and scary time. Light skin makes people think of the safer, daylight hours. But again, the story is likely more complex than this, since humans only began to lose the dark skin of our African ancestors between 11,000 and 19,000 years ago, making race a relatively recent evolutionary development. Skin tone biases are complicated combinations of all kinds of cues, handed down over generations, reinforced, extinguished, and reinforced again. These beliefs can change quickly and exist well below our conscious “radar.”  

Caruso believes that people are making progress at overcoming biases, and explicit cues such as affirmative-action policies and public-service campaigns are having an effect. However, subtle, unconscious cues can still be powerful—because they go unchallenged, they can build over time. “Each time we are exposed to information consistent with a stereotype, the associations that exist in our implicit system are reinforced,” he says. 

Our feelings about race are a function of our political views and the state of the political system, Caruso says. Along with researchers at New York University—Balcetis, Chadly Stern, Shana Cole, and Tessa V. West—Caruso wagered that in an unstable political climate, Americans would gravitate to national values that many find comforting. It’s logical that we’d gravitate to the status quo when we’re feeling unease in the world around us. But in times of stability, they figured, people would feel secure enough to prioritize their own personal values. 

The team tested this hypothesis as it plays out in feelings about biracial candidates, and again, in perception of skin tone. Since previous work has found that Americans of almost all ethnicities, albeit to varying degrees, associate lighter skin with positive traits and darker skin with negative traits, Caruso and colleagues looked at how these “national values” might shift in stable and unstable systems. They were particularly interested in how African Americans’ feelings might shift, since this group has been shown to be less likely to associate lighter skin with positive traits. 

The team had participants read short descriptions of a fictional biracial Department of Education (DOE) candidate and then decide which of three pictures (darkened, lightened, or unchanged) was most representative of him. They variously led the participants to believe that the DOE was going through a period of stability or instability.

In “stable” times, white participants rated lighter pictures of the candidate as being more representative when they agreed with his political beliefs. In contrast, black participants rated the darkened pictures of him as being more representative when they agreed with his beliefs, suggesting that in stable times, we gravitate to personal values. When the participants were led to believe that the system was unstable, however, while the behavior of white participants didn’t change, black participants rated the candidate’s skin tone as lighter when they shared his political beliefs. This dramatic switch suggests a move away from personal values and toward core national values when things are uneasy. This trend also played out in the respective groups’ intentions to vote for the candidate: in unstable times, black participants who perceived the candidate as having a lighter skin tone also expressed more interest in voting for him. There was no variation in the white participants’ voting intentions. 

But what’s particularly notable is the connection between conscious and subconscious processes. When participants were asked explicitly about their values, the same trend emerged, so that “national values became more important when the system was presented as unstable, whereas personal values became more important when the system was presented as stable.” So participants were aware on a conscious level that they preferred the safety of national “status quo” values when they sensed instability, but that they were “freer” to favor their personal beliefs when things seemed more settled. 

What’s going on implicitly? Caruso and his team used the Implicit Association Test (IAT) to see whether people subconsciously associated “light” with positive traits and “dark” with negative traits, as has been shown many times in past research. But here, the team added the “stable” and “unstable” political climates to the mix. In the computerized IAT, participants have to decide very quickly about whether presented images or words are compatible or incompatible. Categories like “dark” or “light” are paired with either positive words (like “rainbow” or “love”) or negative words (“pain” or “evil”). The idea is that depending on the strength of the association (e.g., “rainbow” might fit better with “light”), participants’ reaction times to the pairings will vary. 

It turned out that white participants’ IAT scores didn’t vary when they were told the political climate was stable or unstable, but black participants’ scores did: in stable times, they linked dark skin with positive traits, but in unstable times, they were more likely to link dark skin tone with negative traits. “These associations,” write the authors, “generally lie latent at a nonconscious level, but can guide judgments.” 

The shifts in black participants’ behavior is telling: the political climate affects the perception of race, and it apparently does so explicitly and implicitly. The two processes aren’t always at odds with one another, as Risen showed, but can complement and inform each other—and guide us in choosing political candidates. That is potentially very good news, Caruso says, since it’s likely that as we see more and more examples of high-profile African American individuals in politics and other arenas, the reality that African Americans are increasingly reaching great heights will sink in. Says Caruso, “for example, having Obama in the White House—the implicit system is likely to catch up over time, as the positive associations are reinforced.” 

chartMaking biased basketball foul calls disappear

Obama has spent the better part of the past decade in the White House, putting six years of pressure on any implicit associations that would keep blacks from the highest office in the US. This has challenged some overt beliefs, which Obama has addressed in comments about race, including some about Trayvon Martin, the black, unarmed teenager who was killed by George Zimmerman, a neighborhood watchman, in Florida. "Trayvon Martin could have been me 35 years ago," Obama said after a court controversially found Zimmerman not guilty of manslaughter and second-degree murder.

But we can do more than wait out racism. Explicit knowledge—and in particular, shifts in explicit knowledge—can greatly affect how we process race on all levels of consciousness. Devin G. Pope, associate professor of behavioral science and Robert King Steel Faculty Fellow, took advantage of a natural experiment to illustrate this phenomenon. 

Back in 2007, preliminary results from a study by Brigham Young University’s Joseph Price and the University of Michigan’s Justin Wolfers showed that National Basketball Association (NBA) referees demonstrated racial bias when making foul calls: specifically, when more referees in the three-person referee crew were of a different race from the player, the referee crew called more fouls on that player. The NBA was not pleased with the well-publicized report of bias within its own system, and trashed the paper’s results. The claims of racism by referees came from “a paper that is flat-out wrong in its conclusions,” said NBA president Joel Litvin. Whether awareness of the issue would bring change to the calling of fouls—which is a split-second decision—was, of course, the big question for fans, players, and team owners. 

Pope, Price, and Wolfers took advantage of this heightened awareness to study how calls changed in the following years. In the time between when the paper came out and the media blitz began, the researchers found measurable bias. But they found that after the media coverage, there was a noticeable swing in the number of foul calls that were made: the bias disappeared.  

Its disappearance implies that bias is not a fixed trait, but is instead highly dependent on the environment. Pope points out that referees don’t overcompensate after they’re made aware of the bias: “They actually get it just right,” he says. 

What’s remarkable is that referees are getting it right on a split-second level, which implies that becoming consciously aware of the phenomenon filtered down to the unconscious level. Pope says he’s not surprised that bias is receding in institutions outside of the NBA: banks deciding to be less biased when making loans is a matter of hard, cold facts and conscious thought. But it’s interesting that this particular instance of racism also dissipated in the NBA, where making split-second foul decisions in a basketball game relies on a more subjective, intuitive process.

The effect of revealing racial prejudice

Can we change biases in meaningful ways that will shift and dissolve racist attitudes? Pope says he’s heartened by the idea that simply being aware of the problem is a key to addressing it. “It’s more evidence that if you just say to people, ‘By the way, we’ve got to be careful here,’ it will have an effect on our behavior in meaningful ways.” In a similar vein, a story published in the Journal of General Internal Medicine several years ago showed that doctors assessing hypothetical patients showed a clear race bias, and tended to undertreat African American heart patients. Although the doctors weren’t told they were taking part in a race study, about a quarter had an inkling that was the case—and it was those doctors who did not show racial bias when deciding how to treat. So whether deciding on foul calls or heart medications, awareness
of the race issue is one way to get into our brains and minds and have an effect. 

Implicit cues are just as strong a factor. Bertrand points out that the results about doctors came from a study, not the real world. Using subtler, implicit means to remedy bias may be a powerful accompaniment to conscious awareness of biases. Saying to referees, or the public, “Don’t be racist now,” may only go so far. It may take a complementary approach to attack the unconscious level at which these biases seem to exist. 

The research suggests it will take a marriage of explicit and implicit reminders to undo long-held biases. And as always, one has to be mindful of the difference between what’s already been done and what’s possible in the future. Research by the University of Waterloo’s Richard Eibach and Washington State University’s Joyce Ehrlinger published in the Personality and Social Psychology Bulletin finds that one reason white Americans tend to perceive greater progress toward racial equality than do ethnic minorities is because whites tend to compare the present to the past while minorities tend to compare it to ideal standards that could be achieved in the future. 

If a magazine’s cover shows a black woman as a corporate CEO, for example, a white person may interpret the photo as an indication of how far we have come. A black person might see it as a sign of how far we have to go. “It’s important to keep the goal of equality in mind,” says Risen. The cover may have the biggest impact if it shows the CEO not as a sign of past achievements but as a commitment to that ultimate goal.

Works cited

David Abrams, Marianne Bertrand, and Sendhil Mullainathan, "Do Judges Vary in Their Treatment of Race?" Journal of Legal Studies, June 2012.

Emily Balcetis, Chadly Stern, Shana Cole, Tessa V. West, and Eugene M. Caruso, “Government Instability Shifts Representations of Political Candidates’ Skin Tone,” Working paper, 2014. 

Marianne Bertrand, Dolly Chugh, and Sendhil Mullainathan, “Implicit Discrimination,” American Economic Review, May 2005. 

Marianne Bertrand and Sendhil Mullainathan, “Are Emily and Greg More Employable Than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination,” American Economic Review, September 2004. Chart reprinted with permission from the American Economic Association. Copyright 2004.

Eugene M. Caruso, Nicole L. Mead, and Emily Balcetis, “Political Partisanship Influences Perception of Biracial Candidates’ Skin Tone,” Proceedings of the National Academy of Sciences, May 2009. Chart reprinted with permission from Proceedings of the National Academy of Sciences. Copyright 2009. 

Clayton Critcher and Jane L. Risen, “If He Can Do It, So Can They: Exposure to Counterstereotypically Successful Exemplars Prompts Automatic Inferences,” Journal of Personality and Social Psychology, March 2014. Chart reprinted with permission from the American Psychological Association. Copyright 2014. 

Richard P. Eibach and Joyce Ehrlinger, “’Keep Your Eyes on the Prize’: Reference Points and Group Differences in Assessing Progress towards Equality,” Personality and Social Psychology Bulletin, January 2006.

Alexander R. Green, Dana R. Carney, Daniel J. Pallin, Long H. Ngo, Kristal L. Raymond, Lisa I. Iezzoni, and Mahzarin R. Banaji, “Implicit Bias among Physicians and Its Prediction of Thrombolysis Decisions for Black and White Patients,” Journal of General Internal Medicine, September 2007. 

Devin G. Pope, Joseph Price, and Justin Wolfers, “Awareness Reduces Racial Bias,” NBER working paper, December 2013.