A Nobel Laureate on the Limits of Evidence-Based Policy
Lars Peter Hansen and Kevin M. Murphy discuss how data can inform policymaking.
A Nobel Laureate on the Limits of Evidence-Based PolicyDustin Whitehead
Narrator: The tone of your skin, the color of your hair, the shape of your eyes: how people judge physical characteristics has shaped history. At times it has sparked oppression, violence, and heartache. Some say racism is in the process of being consigned to history. After all, in the United States, institutionalized slavery is a thing of the past. The law guarantees racial equality, and people from minority groups occupy positions of power and influence.
But research conducted at the University of Chicago Booth School of Business suggests racism is still a powerful force in society. Take a look at how people mingle in your favorite bar, for example.
Emir Kamenica: In order to identify who wants to be with whom rather than who ends up with whom, we ran the speed-dating experiment at Columbia University. We organized it for research purposes so we could observe yes and no decision—Do I want to go out with this person or not?—for every possible pairing. So every person would say yes or no to everybody else, so we would see not just final outcome, but for every person, we knew exactly what was the set of people they met that they’d like to go out with, versus not.
Woman: Would you date me?
Man: Am I attractive?
Man: Would you date me?
Emir Kamenica: It’s called speed dating because these are not very long encounters. They can’t be if you’re gonna meet 22 people. In our instances, they were four-minute dates. In one of these papers, we focused explicitly on racial preferences for romantic partners. There, we’re simply looking at, for white, Black, and East Asian, what kind of attributes they seek in their partners, and whatever the attributes might be, the partner be of the same race.
Man: I would date a Black girl.
Woman: I’ll date anyone.
Woman: Yes, I would date a white guy.
Emir Kamenica: The strongest pattern we find is that women seem to be much more sensitive to the race of the partner than men are.
Woman: I have not dated a Black man.
Emir Kamenica: We asked all of the participants about where they grew up, the zip code where they grew up, and we can match that to, say, the income of their zip code. We can match that to the racial composition of the zip code where they grew up.
But one thing we find is that growing up around people of a particular race does not diminish the dislike of that race compared to your own. If anything, in fact, the same race preferences tend to be somewhat stronger for those who grew up surrounded by people of another race. For example, if we look at Black women who grew up in neighborhoods where there are more white, they are less likely to say yes to a white man compared to Black women that grew up in a neighborhood that were fewer white people.
Certain attributes seem correlated with not having a strong same-race preference. One is age. Older people seem to be less sensitive to whether the potential partners are the same race or not.
And then another finding is that attractive people are a lot less sensitive to race. So we have lots of good measures of how attractive these people are. We’ve got RAs [research assistants] that walked around and did a subjective measures of attractiveness. We know how attractive other participants found them, so physically attractive people seem to be noble, in a sense, that they don’t seem to seek partners of the same race.
(Somber music) You know, if you look back to the 2000 census, which was the last census at the time we did our study, I believe it’s something like 94 percent of all Black people in the US who were married are married to another Black person. This level of segregation, marital segregation, is really quite high compared to all kind of other forms of segregation do you ask.
If you look at, you know, segregation when it comes to who are neighbors of a typical individual, of course, there’s plenty of geographic segregation in the US. But we never see numbers that are quite as extreme as the numbers we’ve seen when it comes to marriages, where really the interracial marriages are still quite rare. This is a different sort of a choice. This is a very intimate choice. This is a choice about whom to be intimate with.
Narrator: The results suggest people who are dating gravitate to others of the same race. It’s possible to see what’s happening in the bar, but the psychological process is less obvious. Eugene Caruso, a behavioral scientist, has found a link between how people feel about a person and what they see.
Eugene M. Caruso: Trying to understand how people kind of on opposing sides of different issues can come to see information very differently. So even when people are looking at the exact same information, they often come to very different kind of conclusions about, say, its merit or its worth.
What we did was we identified some biracial political candidates, that is, where they were sort of ambiguity over the actual color of their skin tone, and we presented pictures of these candidates to research participants. In one study, for instance, we gave participants pictures of Barack Obama. Now, unbeknownst to participants, for some of the photographs we artificially lightened his skin tone and for some of the photographs we artificially darkened his skin tone. One original version, one lightened version, and one darkened version, so for each photograph we had the participants rate how well this photograph represented who the candidate really was.
Woman: What do you see?
Woman: What do you see?
Man: What do you see?
Eugene M. Caruso: So we asked them a series of other questions, such as their own political ideology and their political beliefs, how likely they were to vote for a particular candidate. What we found was that participants whose own political ideology matched that of the candidate they were evaluating tended to rate the lightened photos as more representative of who he really is. So, for Barack Obama, liberals rated the lightened versions of his skin tone as representing him better than the darkened photos, whereas political conservatives showed the opposite effect.
Narrator: The study participants were primarily white. When Caruso and his colleagues conducted a similar test with a more diverse population, they found a more nuanced story.
Eugene M. Caruso: We manipulated the perceived stability of the political system by simply telling participants about the state of affairs for the government in which the candidate they were evaluating was operating. And so we would describe a time of relative, you know, kind of stability or a time of relative instability and uncertainty and chaos to basically activate the idea in the participants’ minds that the candidate is either operating in a system that’s relatively stable or relatively instable.
Man: Who do you see?
Woman: Who do you see?
Man: Who do you see?
Eugene M. Caruso: When the political system was described as being unstable, white participants and Black participants both showed the same effect. That is, when the candidate they were evaluating agreed with them politically, they were more likely to rate the lightened photos as more representative of who he really is than the darkened photos, but in times of system instability, Black participants actually showed kind of a different affect, whereby they were not faster to associate lightness with good than they were darkness with good.
So what we think is going on is that when the system is relatively unstable, people tend to band together and prioritize national values above all else, and partisan differences tend to be kind of, you know, brushed aside. However, in times of relative stability, we think that that enables people to act more on their own personal values and their own personal ideological beliefs.
And so we think that we find a similar affect with people’s race. That is, in times of stability, Black participants are basically enabled to act on their own personal values that associate their own race with positivity, and candidates who may have a darker skin tone are seen as more positive than those with a lighter skin tone. And we think that participants aren’t aware of these associations that they have, or certainly are not aware that their ideology can affect how they see political candidates, and so it suggests that the barriers to overcoming some of these differences may be pretty high.
Narrator: Caruso’s research suggests people sometimes associate white with good, and black with bad. Recognizing this could be helpful, but it presents its own challenges.
In November 2008, tens of thousands of people took to the streets of Chicago as Barack Obama was elected the first African American president of the United States. While many saw this as a step in the right direction, few would explicitly say that it indicated racism had been eliminated. But [Chicago Booth’s] Jane Risen suggests that what we say about racism may not be the same as what we really think.
Jane L. Risen: What we were looking for was to see how people thought about race relations after being incidentally exposed to exemplars that were counter stereotypically successful in white domain.
When I say incidentally exposed, what I mean is that we present people with two different tasks. They think that they’re doing two separate studies. In the first study, they see what’s called a sort of a celebrity recognition task, and there are six different questions, and what we change between our experimental condition and our control condition is whether one of the exemplars is a counter stereotypically successful African American.
So it might be you have a set of white controls and then one of the people in the celebrity quiz would be Oprah, where, in the control condition, it would be Barbara Walters. They see these examples in the celebrity recognition task. They see a picture so that they know who the race is, even if they don’t know who this person is, and then they think that that’s done. We’re just pretesting to find out how much you know about celebrities.
Then, in the second task, they’re asked about race relations. Basically, what we find in lots of different ways is that once you’re incidentally exposed to an example, like Tony Morrison or Ruth Simmons, who is the president of Brown, or Obama, or Oprah, afterwards then people say that there’s actually, they sort of deny racism existing. So they say that currently we are in a state with more equality and opportunity, and to the extent that they are asked, sort of, what is the cause of anything that’s unequal, people tend to blame the Black community more rather than blaming, sort of, persistent racism. So when we ask people explicitly, “To what extent does the existence of Obama, or does the existence of Oprah, the existence of any of these guys, to what extent does their existence make you believe that there is less racism?” People say no.
So people sort of know that this is, you know, that you can’t reason from a single example, right? A single example is just not informative for the status of race relations broadly, but when they’re exposed to just one example, they show this inference unintentionally.
If we think about these symbols as a form of commitment to our goal of, sort of, egalitarianism, then that can be very helpful, and so I think if you could get people generally focused on this ideal of what we’re trying to do and feel committed to that ideal, then one example of success would no longer be a signal of things are already done. This is still a work in progress. This is not something that’s been completed.
Narrator: While Risen’s experiments reveal our deep reviews about the persistence of racism, other research might actually suggest ways to combat prejudice. Simply being aware of racist behavior may help to address it. That’s what [Booth’s] Devin Pope, a behavioral economist, found when he got interested in a study by two economists, [Brigham Young’s] Joseph Price and [University of Michigan’s] Justin Wolfers, that looked at foul calls made by NBA referees. A study that received a lot of media attention.
Devin G. Pope: They found that if the referee crew was primarily Black, then white players receive more foul calls relative to Black players. And when the referee crew was primarily white, then Black players received more foul calls relative to white players.
In our follow-up study, we were able to look at data from 2003 to 2006—so this is data that’s subsequent to the original paper’s data, but before the media attention—to see if after the media attention, the effect continues to persist. And we see it goes away completely. So we find no evidence of racial bias after 2007, when the original paper received a lot of the media attention.
We can’t say with a lot of certainty exactly what made the bias go away. We see that the bias does go away, and it wasn’t due to the fact that the NBA fired the referees that our results suggested had the largest amount of bias. The NBA claimed that there were no institutional changes. They claim that they had never even spoke with the referees about this, so it appears that these referees just, on their own, because they are aware of their bias, they started to eliminate that bias in their own behavior. And it’s not clear exactly what the mechanism is, but it appears that somehow, by simply being aware of it, they were able to eliminate it.
Woman: I’m not racist.
Man: I’m not racist.
Man: I’m not racist.
Man: I thought I wasn’t.
Devin G. Pope: So the findings that we have are at one particular case. It’s in the NBA. It’s a situation where these referees are making decisions that are gonna be seen by others and can be tested by others to see if they’re being unbiased. Making people aware of their biases, in this particular situation, we have shown works, but we’re not exactly sure how this would generalize to other situations. It’s possible that the referees feel a lot of pressure to do things that are socially acceptable in this particular domain, and that the awareness and the publicity and everything is enough to push them over the edge.
Or it might not be enough in other domains, or other people, when they’re made aware of their biases but it’s not so publicly done, may not see the same effect. So there’s clearly limitations to what we can learn from this study, but at least it gives us one point of evidence where awareness can reduce racial bias.
Man: Am I a racist?
Woman: Am I a racist?
Woman: Are we racist?
Devin G. Pope: You could imagine other situations where employers, when they’re shown that they actually maybe are racially biased in their hiring decisions, would be grateful for that knowledge and would try to improve. And, yet, they just simply weren’t aware of the fact that they were discriminating based on race.
Man: Are we racist?
Man: Are we racist?
Man: Are we racist?
Devin G. Pope: Having people get to know others that they are discriminating against, the mere exposure to others can reduce discrimination. But this is another one, making people aware of their own discrimination can help them want to do better.
Narrator: Collecting data about how we think and act reveals patterns about our behavior, and it may also help us address the bigger issue: racial prejudice that persists in spite of the progress that we’ve made.
Jane L. Risen: It’s not enough to just get rid of explicit bias. We also, you know, for this true goal of equality, we wanna get rid of the implicit bias, too, and that’s a longer trail, I think.
Man: Have we overcome?
Woman: Have we overcome?
Man: Have we overcome?
Child: Have we overcome?
Man: Have we overcome?
Man: Have we overcome? Not yet.
Lars Peter Hansen and Kevin M. Murphy discuss how data can inform policymaking.
A Nobel Laureate on the Limits of Evidence-Based PolicyThree experts discuss the sources of income inequality.
Why Are the Very Rich Getting Even Richer?Job loss, divorce, and other life events far outweigh negative equity.
What’s Actually behind Most Mortgage DefaultsYour Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.