I took the test a second time, and then a third time, and then a fourth time, hoping that the awful feeling of bias would go away. It made no difference. It turns out that more than 80 percent of all those who have ever aken the test end up having pro-white associations, meaning that it takes them measurably longer to complete answers when they are required to put good words into the “Black” category than when they are required to link bad things with black people. I didn’t do quite so badly. On the Race IAT, I was rated as having a “moderate automatic preference for whites.” But then again, I’m half black. (My mother is Jamaican.)

So what does this mean? Does this mean I’m a racist, a self-hating black person? Not exactly. What it means is that our attitudes toward things like race or gender operate on two levels. First of all, we have our conscious attitudes. This is what we choose to believe. These are our stated values, which we use to direct our behavior deliberately. The apartheid policies of South Africa or the laws in the American South that made it difficult for African Americans to vote are manifestations of conscious discrimination, and when we talk about racism or the fight for civil rights, this is the kind of discrimination that we usually refer to. But the IAT measures something else. It measures our second level of attitude, our racial attitude on an unconscious level—the immediate, automatic associations that tumble out before we’ve even had time to think. We don’t deliberately choose our unconscious attitudes. And as I wrote about in the first chapter, we may not even be aware of them. The giant computer that is our unconscious silently crunches all the data it can from the experiences we’ve had, the people we’ve met, the lessons we’ve learned, the books we’ve read, the movies we’ve seen, and so on, and it forms an opinion. That’s what is coming out in the IAT.

The disturbing thing about the test is that it shows that our unconscious attitudes may be utterly incompatible with our stated conscious values. As it turns out, for example, of the fifty thousand African Americans who have taken the Race IAT so far, about half of them, like me, have stronger associations with whites than with blacks. How could we not? We live in North America, where we are surrounded every day by cultural messages linking white with good. “You don’t choose to make positive associations with the dominant group,” says Mahzarin Banaji, who teaches psychology at Harvard University and is one of the leaders in IAT research. “But you are required to. All around you, that group is being paired with good things. You open the newspaper and you turn on the television, and you can’t escape it.”

The IAT is more than just an abstract measure of attitudes. It’s also a powerful predictor of how we act in certain kinds of spontaneous situations. If you have a strongly pro-white pattern of associations, for example, there is evidence that that will affect the way you behave in the presence of a black person. It’s not going to affect what you’ll choose to say or feel or do. In all likelihood, you won’t be aware that you’re behaving any differently than you would around a white person. But chances are you’ll lean forward a little less, turn away slightly from him or her, close your body a bit, be a bit less expressive, maintain less eye contact, stand a little farther away, smile a lot less, hesitate and stumble over your words a bit more, laugh at jokes a bit less. Does that matter? Of course it does. Suppose the conversation is a job interview. And suppose the applicant is a black man. He’s going to pick up on that uncertainty and distance, and that may well make him a little less certain of himself, a little less confident, and a little less friendly. And what will you think then? You may well get a gut feeling that the applicant doesn’t really have what it takes, or maybe that he is a bit standoffish, or maybe that he doesn’t really want the job. What this unconscious first impression will do, in other words, is throw the interview hopelessly off course.

Or what if the person you are interviewing is tall? I’m sure that on a conscious level we don’t think that we treat tall people any differently from how we treat short people. But there’s plenty of evidence to suggest that height—particularly in men—does trigger a certain set of very positive unconscious associations. I polled about half of the companies on the Fortune 500 list—the list of the largest corporations in the United States—asking each company questions about its CEO. Overwhelmingly, the heads of big companies are, as I’m sure comes as no surprise to anyone, white men, which undoubtedly reflects some kind of implicit bias. But they are also almost all tall: in my sample, I found that on average, male CEOs were just a shade under six feet tall. Given that the average American male is five foot nine, that means that CEOs as a group have about three inches on the rest of their sex. But this statistic actually understates the matter. In the U.S. population, about 14.5 percent of all men are six feet or taller. Among CEOs of Fortune 500 companies, that number is 58 percent. Even more striking, in the general American population, 3.9 percent of adult men are six foot two or taller. Among my CEO sample, almost a third were six foot two or taller.

The lack of women or minorities among the top executive ranks at least has a plausible explanation. For years, for a number of reasons having to do with discrimination and cultural patterns, there simply weren’t a lot of women and minorities entering the management ranks of American corporations. So, today, when boards of directors look for people with the necessary experience to be candidates for top positions, they can argue somewhat plausibly that there aren’t a lot of women and minorities in the executive pipeline. But this is not true of short people. It is possible to staff a large company entirely with white males, but it is not possible to staff a large company without short people. There simply aren’t enough tall people to go around. Yet few of those short people ever make it into the executive suite. Of the tens of millions of American men below five foot six, a grand total of ten in my sample have reached the level of CEO, which says that being short is probably as much of a handicap to corporate success as being a woman or an African American. (The grand exception to all of these trends is American Express CEO Kenneth Chenault, who is both on the short side—five foot nine—and black. He must be a remarkable man to have overcome two Warren Harding errors.)

Is this a deliberate prejudice? Of course not. No one ever says dismissively of a potential CEO candidate that he’s too short. This is quite clearly the kind of unconscious bias that the IAT picks up on. Most of us, in ways that we are not entirely aware of, automatically associate leadership ability with imposing physical stature. We have a sense of what a leader is supposed to look like, and that stereotype is so powerful that when someone fits it, we simply become blind to other considerations. And this isn’t confined to the executive suite. Not long ago, researchers who analyzed the data from four large research studies that had followed thousands of people from birth to adulthood calculated that when corrected for such variables as age and gender and weight, an inch of height is worth $789 a year in salary. That means that a person who is six feet tall but otherwise identical to someone who is five foot five will make on average $5,525 more per year. As Timothy Judge, one of the authors of the height-salary study, points out: “If you take this over the course of a 30-year career and compound it, we’re talking about a tall person enjoying literally hundreds of thousands of dollars of earnings advantage.” Have you ever wondered why so many mediocre people find their way into positions of authority in companies and organizations? It’s because when it comes to even the most important positions, our selection decisions are a good deal less rational than we think. We see a tall person and we swoon.