IRA FLATOW, HOST:
This is SCIENCE FRIDAY. I'm Ira Flatow. If the polls are a good indicator, the economy, jobs, the deficit, health care and education are likely to be the issues that weigh heavily on voters' minds when they head to the polls in November. But researchers say there may be another factor that influences the presidential vote this election cycle, and that's racial attitudes.
Four years after Americans elected their first black president, a new study shows that it is premature to count race out as a factor. Dr. Tony Greenwald was lead investigator for the study and is a professor of psychology at the University of Washington. He joins us from KUOW in Seattle. Welcome to SCIENCE FRIDAY.
ANTHONY GREENWALD: Hi, Ira, thanks.
FLATOW: You're welcome. Why do a study on racial attitudes in 2012, after we've had, you know, the election of President Obama in 2008? You wouldn't think this is a big issue anymore.
GREENWALD: Well, it was an open question whether there's - race attitudes are still affecting Americans' political attitudes or not. We certainly found that they were in 2008. Obama won, but he actually won despite race biases that were opposing him. Obama lost the white vote by 55 to 45 percent, which is a rather large margin.
So we did a similar study, much larger this year, just to see what was happening.
FLATOW: And what did you find out?
GREENWALD: We found that the level of race bias indicating preference, and this was in the Republican primaries, we were asking about support for the different Republican candidates and comparing them with Obama, we found that this was predicted by race attitudes of several kinds.
We measured race attitudes in a few different ways, such as asking people explicitly what they thought of black and white, and a lot of people in our sample thought that black and white are just equal, they have no racial preference. But we also used the implicit association test, which gives us a measure of implicit bias, we sometimes call it automatic white preference, and that's something that many more people have.
FLATOW: So you found that even though people thought they were unbiased, when you tested them, they actually showed bias.
GREENWALD: Yes, and we looked especially at people who are most likely to think they are lacking in bias, such as the self-described egalitarians, also political liberals. Our sample, being an Internet sample, is not surprising that it was a majority of liberals. And we found that the implicit race attitudes of these liberals and the egalitarians predicted their warmth toward Republican candidates.
FLATOW: So in other words, many white and non-white voters will unconsciously vote against President Obama because of his race, but when asked, they'll give a reason other than race.
GREENWALD: That's very likely. This is very hard to tease apart, and our method works by measuring things that you can't really ask people about because they don't know. And we think it's one of a number of factors that play into determining vote. But it's a strong enough factor so that it can actually affect a significant percentage of the vote.
FLATOW: Among Republican candidates, when you polled them, did you see similar racial attitudes towards Herman Cain?
GREENWALD: Cain was very interesting because Democrats actually, with racial bias, either explicit meaning they were not egalitarian or implicit, preferred Cain, preferred Cain to Obama. Those who opposed Obama - excuse me, the preference for Cain was predicted by their racial attitudes meaning people who were opposing Obama on the basis of race actually indicated that they preferred Cain, which indicates that they're strong enough against Obama that they would even support Cain.
But when we compared Cain against other Republican candidates and looked just at people who were supporting Republican candidate, the racial attitudes predicted, strongly predicted preference for the white candidates.
FLATOW: But the test itself isn't a measure of racism.
GREENWALD: It is definitely not a measure of racism. We are very clear to emphasize that this measure something that is perhaps unconscious, in any case most people are unaware of it, and it is possessed by a large of the American population, people who do not at all think of themselves as racist. And we don't think they're lying to us when they say they have egalitarian racial beliefs.
FLATOW: Why is it that they're out of touch with their subconscious? This is too big a question to ask, it may take years to answer for any, you know, any person. But do you have any guesses on that?
GREENWALD: Often their explicit race beliefs have been shaped by factors other than the culture as a whole, but our media, entertainment, news, literature, history are filled with negative references to African-Americans compared to white Americans. And those, that environment, the sort of intellectual air we breathe is so full of negative associations with African-Americans that this gets picked up by the brain and stored as these associations that we can measure.
But people aren't actually aware that they have this. They often explicitly reject it. They certainly don't want to have it. But nevertheless, it can act on them, and it can affect their behavior. It can produce discomfort in interracial interactions, and that's something that all by itself is likely to produce some unintended discrimination.
FLATOW: But a lot of people would say, you know, people still have racial biases. Why even study that? We know that there are racial biases around.
GREENWALD: Well, the history of race bias in the United States has been very interesting. In the last 50 years or so, partly due to the legalities of civil rights law, overt racial bias has dropped way down. Very few people indicate any kind of opposition to desegregation in schooling or residential integration.
But at the same time, racism survives in forms that are measured indirectly, including by our test and by some other tests that are more subtle, trying to measure things that are called modern racism or symbolic racism. So racism has not disappeared, but it has taken a very different form, such a different form that we no longer want to use the word racism for it.
We call it implicit race bias or implicit race attitude.
FLATOW: And now that we have Mitt Romney as the presumptive Republican nominee, are you going to continue your survey?
GREENWALD: We - yes, we are continuing by removing the other Republicans from our survey, focusing on Romney and Obama and trying to get at factors that will allow us to understand more why we observe the correlations that we consistently observe.
FLATOW: Tell us about the test. What is the implicit association test? How does that work?
GREENWALD: This is a procedure that's sort of fun to play with. You can take it yourself on the Internet. It's a procedure that involves four steps. And first you're asked to rapidly classify white and black faces by pressing a left-side computer key for black faces or a right key for white faces. That's an easy test.
Second, same type of test, but now you're pressing the left key for unpleasant words, words like hurt or failure, and the right key for pleasant words for words like joy and love. That's an easy task, too.
Then it begins to get more difficult. We combine the tasks, the third step. Now the left key is for either white faces or unpleasant words, and the right key is for either black faces or pleasant words. This combination of two easy tasks can be quite difficult.
When I first did this, which was in the mid-1990s, I was amazed to discover how slowly I was going at the task. I was also making errors. Then there's the fourth step: We switch the sides of the faces. Now it's the black faces and unpleasant words on the left, the white faces and pleasant words on the right. This turned out to be surprisingly easy for me. I could go fast, few or no errors.
This is how I discovered that I have this implicit race bias myself, and as we continued to study this, I found that I was far from alone. About 75 percent of the subjects in our research and the people who take this test on the Web show similar pattern to what I showed.
FLATOW: And we actually have a link to how to take the test up on our website at sciencefriday.com. People can try it out for themselves. What other things can you measure with this test besides bias in that way?
GREENWALD: It's been very useful in measuring gender stereotypes. For example, it shows that there are strong stereotypes that associate male with science, female with arts, and we found that this is correlated with different countries. The strength of this implicit stereotype, is what we call it, predicts male-female differences in performance on standardized tests that are administered every few years.
There's a gender career stereotype that associates women with family, men with careers, and you'd think that this is something that men would show, but women would reject. Actually, the implicit association test shows that women have this stereotype, this implicit stereotype, even a little more strongly than men. And we found in research, others have found in research, that this stereotype is associated with women experiencing difficulty in the workplace.
It's sort of a force inside their head that gives them a source of conflict or a feeling of discomfort in career situations.
FLATOW: Speaking of inside your head, can you tell anything about people's mental health from the test?
GREENWALD: There is now a very active research by clinical psychologists on this, and there is a site that's available called Project Implicit Mental Health at which people can take these tests. I think it can be found on the same page that you will have a link to, but I'm not sure about that.
And one of the uses of this test has been to predict suicide risk. This has been very surprising research that an implicit association test actually predicted risk of reattempting suicide more effectively than self-report, characterizations of the likelihood of recurrence.
FLATOW: And one quick question, about 30 seconds, if you found racial bias with this test, how is this test not a predictor of racism or measure racism?
GREENWALD: It's correlated with overt racism, but it's very different, and the difference is that people are unaware of this. They don't want it, very often. They certainly don't endorse it, and it's what clinical psychologists or psychiatrists call a dissociation. It's like two separate things in the head, one egalitarian and open, the other biased.
FLATOW: All right, Dr. Greenwald, thank you very much.
GREENWALD: Thank you, Ira.
FLATOW: Tony Greenwald, psychologist, professor of psychology at the University of Washington. We're going to take a break, and afterwards we're going to come back and ask the candidates in a science debate - what would you ask them about science? Stay with us. We'll be right back. I'm Ira Flatow. This is SCIENCE FRIDAY from NPR. Transcript provided by NPR, Copyright NPR.