Interview with Mahzarin Banaji

Everyone is biased: Harvard professor’s work reveals we barely know our own minds
By Carolyn Y. Johnson, Boston Globe Staff

Mahzarin R. Banaji was starting out as an assistant professor of psychology at Yale University in the late 1980s, at a time when women professors were scarce enough that administrators eager to offer a class on the psychology of gender turned to her. Banaji had no expertise in the area; her research focused on memory. But she said she would do it, and she quickly found herself inhabiting the overlapping worlds of gender studies and psychology.

Banaji was fascinated by a memory study by psychologist Larry Jacoby. He had asked people to read a list of names from the phone book, such as “Sebastian Weisdorf”, and rate how easy they were to pronounce. A day later, those same people were handed a list of names that included famous people, others from the phone book, and some names from the list they had read the day before. Asked which were famous people, the study participants incorrectly classified Sebastian Weisdorf and others, whose names they had learned just the day before, as famous.

What, Banaji wondered, would happen if the name was Susannah Weisdorf? Would this same benefit, of becoming famous overnight, accrue to women? She did the test and found that female names were far less likely to achieve fame in the same way. When she grilled participants later, to try and figure out what could lie behind the discrepancy, she was struck by one thing: it occurred to no one that gender might be a factor.

That study was a seed, which grew into an idea in psychology that has become transformative: everyone carries with them implicit biases that may change how people perceive or interact with others. Doctors, judges, police officers, teachers—even Banaji herself—are all subject to these biases, which can lead people to inadvertently act in ways that may be discriminatory or are influenced by stereotypes that people would consciously reject.

Today, millions of people have taken the Implicit Association Test that Banaji, Anthony G. Greenwald of the University of Washington, and Brian Nosek of the University of Virginia developed together. The tests often reveal distressing things about how our unconscious minds work. Most famously, people presented with a computer screen that asks them to press different keys to associate adjectives with faces more quickly associate positive adjectives with white faces than with black faces.

Banaji, now a psychology professor at Harvard University, has with Greenwald written a book, “Blindspot: Hidden Biases of Good People,” describing the origins and development of the influential idea. I spoke with Banaji by telephone.

What are your own biases?

A: The first experience I had with the IAT [Implicit Association Test] was very disconcerting. I mean, I was quite certain that if I took a black-white test, to associate black and white with good and bad, I was certain that I would come out associating black with good. … I’ve spent a lot of my life thinking about these issues. I am, myself, an immigrant. I was aware of the history of black-white relations; I had strived in my own life to practice what I believe. And among my peers, I have the reputation of someone who understands these issues and cares about them. I should certainly not have trouble.

So when I took the test … it was stunning for me to discover that my hands were literally frozen when I had to associate black with good. It’s like I couldn’t find the key on the keyboard, and doing the other version, the white-good, black-bad version was trivial.

So the first thought that I had was: “Something’s wrong with this test.” Three seconds later, it sunk in that this test was telling me something so important that it would require a re-evaluation of my mind, not of the test. That’s where it sort of began.

When was the term “implicit bias” coined?

I had been working in the area of memory, [and there was a new notion of] implicit memory. They had, in the ‘80s really experienced a little revolution in their field, where all of a sudden they were discovering that memory has been measured for 100 years by psychologists who asked people questions: What did you have for breakfast? What do you remember from when you were five years old? All of those require conscious memory. There are people, amnesic patients, who have no such recollection. They can’t remember what happened that morning; they can’t remember the faces of their doctors of 10 years. But they did have memory for some things, they just didn’t have conscious memory. The examples were if you told an amnesic patient a joke many times, they could not remember the punch line, but they would begin to not find the joke funny.

I think if we did anything that was unique, it was to take that way of thinking about memory—that memory may come in two forms, conscious memory and implicit memory—and ask, is it true not only for memory but our implicit beliefs, our likes and our dislikes?

I remember writing an e-mail to a colleague of mine … late at night and I said, “What do you think about calling this stuff ‘implicit attitudes’?” He wrote back at 2 a.m. in this morning: “It has an auspicious ring to it. I would use it.”

If we all have them, what can people do about these biases?

A: I would say we should not be naive about how easily we can change them. On the other hand, there are studies that demonstrate that you can at least produce shifts. There are studies that show when you are confronted with an opposite association from the one you are usually used to seeing, it will shift you in some way. One of my students did a study looking at women who went to a single-sex school vs. women who went to a co-ed school. She measured them at the beginning and found them roughly the same in the gender stereotypes. … Four years later, they look different. Women in the single-sex school show weakened gender stereotypes. … It’s literally the number of women teachers you’ve had.

At its heart, the test is telling us something about ourselves that is not palatable. While that leads us to reject this message, it also keeps us coming back to it, and once you know that, it opens up a whole new universe of decisions that can be made quite differently.

You go have a cholesterol test or a hypertension test and it tells you: you have something in your body of which you could not have been aware. It’s not your fault, but once you know, a smart person will do something differently.

What has been the most surprising result over the years?

There is a series of studies that stand out in my mind because of how shocking the result was. This is the study when we ask, “Who is American?” We did it by giving people an IAT in which they saw faces of Americans: some are of Asian origin, Americans of European origin, but the test takers were told they were all born and raised in the U.S. We had the test set up so they had to [match the faces] with symbols that were obviously American symbols or symbols of foreign countries. Lo and behold, you get this result that’s not completely unexpected. People have a quicker association between European faces and American symbols than Asian faces and American symbols. You could say, Asians came to this country later and … the category American means you’ve been here a very, very long time.

If that’s the case, use native American faces and European American faces and now let’s associate them with American symbols and foreign symbols. Because things like the dollar bill and Washington monument were created by Europeans, let’s use natural scenes: Niagara Falls and the Grand Canyon. Still, we find European-Americans are associated more strongly with American symbols. So, to be American is to be white.

And we took famous Asian-Americans, like Connie Chung and Michael Chang and Kristi Yamaguchi and then we picked white foreigners: Hugh Grant, Katarina Witt, Gerard Depardieu. You know they’re not American. Now, associate them with American symbols and foreign symbols. It was easier to associate Hugh Grant with American symbols than Connie Chung; that shows how deeply the category “American” is white.

The reason I especially like that result is that in the first Obama election and since then, the issue has come up about these “birthers,” and I think what we captured there was a little bit of a birther in all of us. I think this is where conscious attitudes matter. You and I say, “I consciously know Barack Obama was born in this country, and I believe this because the evidence is there.” For some people who we might write off as the lunatic fringe, the association to be American is to be white. I can see for them that feels true.

It’s when an event like the birther issue happens, that’s when you realize what you’re tapping into is not just a laboratory demonstration.