Tajfel and Turner called their insight "social identity theory," which has proved valuable for understanding how prejudices develop. Given even the slenderest of criteria, we naturally split people into two groups—an "in-group" and an "out-group." The categories can be of geopolitical importance—nationality, religion, race, language—or they can be as seemingly inconsequential as handedness, hair color or even height.
Once the division is made, the inferences and projections begin to occur. For one, we tend to think more highly of people in the in-group than those in the out-group, a belief based only on group identity. Also, a person tends to feel that others in the in-group are similar to one's self in ways that—although stereotypical—may have little to do with the original criteria used to split the groups. Someone with glasses may believe that other people who wear glasses are more voracious readers—even more intelligent—than those who don't, in spite of the fact that all he really knows is that they don't see very well. On the other hand, people in the out-group are believed to be less distinct and less complex than are cohorts in the in-group.
Although Tajfel and Turner found that identity and categorization were the root cause of social bias, other researchers have tried to find evolutionary explanations for discrimination. After all, in the distant past, people who shared cultural similarities were found to be more genetically related than those who did not. Therefore, favoring the in-group was a way of helping perpetuate one's genes. Evolutionary explanations seem appealing, since they rely on the simplest biological urges to drive complicated behavior. But this fact also makes them hard to prove. Ironically, there is ample evidence backing up the "softer" science behind social identity theory.
Hidden Bias
Not many of us will admit to having strong racist or xenophobic biases. Even in cases where bias becomes public debate—such as the profiling of Arab Muslims at airport-security screenings—proponents of prejudice claim that they are merely promoting common sense. That reluctance to admit to bias makes the issue tricky to study.
To get around this problem, psychologists Anthony Greenwald, of the University of Washington in Seattle, and Mahzarin Banaji, of Harvard, developed the Implicit Association Test. The IAT is a simple test that measures reaction time: The subject sees various words or images projected on a screen, then classifies the images into one of two groups by pressing buttons. The words and images need not be racial or ethnic in nature—one group of researchers tested attitudes toward presidential candidates. The string of images is interspersed with words having either pleasant or unpleasant connotations, then the participant must group the words and images in various ways—Democrats are placed with unpleasant words, for instance.
The differences in reaction time are small but telling. Again and again, researchers found that subjects readily tie in-group images with pleasant words and out-group images with unpleasant words. One study compares such groups as whites and blacks, Jews and Christians, and young people and old people. And researchers found that if you identify yourself in one group, it's easier to pair images of that group with pleasant words—and easier to pair the opposite group with unpleasant imagery. This reveals the underlying biases and enables us to study how quickly they can form.
Really though, we need to know very little about a person to discriminate against him. One of the authors of this story, psychologist Margo Monteith, performed an IAT experiment comparing attitudes toward two sets of made-up names; one set was supposedly "American," the other from the fictitious country of Marisat. Even though the subjects knew nothing about Marisat, they showed a consistent bias against it.
While this type of research may seem out in left field, other work may have more "real-world" applications. The Southern Poverty Law Center runs a Web version of the IAT that measures biases based on race, age and gender. Its survey has, for instance, found that respondents are far more likely to associate European faces, rather than Asian faces, with so-called American images. The implication being that Asians are seen as less "American" than Caucasians.
Similarly, Harvard's Banaji has studied the attitudes of people who favor the racial profiling of Arab Muslims to deter terrorism, and her results run contrary to the belief that such profiling is not driven by xenophobic fears. "We show that those who endorse racial profiling also score high on both explicit and implicit measures of prejudice toward Arab Muslims," Banaji says. "Endorsement of profiling is an indicator of level of prejudice."
Beyond Xenophobia
If categorization and bias come so easily, are people doomed to xenophobia and racism? It's pretty clear that we are susceptible to prejudice and that there is an unconscious desire to divide the world into "us" and "them." Fortunately, however, research also shows that prejudices are fluid and that when we become conscious of our biases we can take active—and successful—steps to combat them.
Researchers have long known that when observing racially mixed groups, people are more likely to confuse the identity of two black individuals or two white ones, rather than a white with a black. But Leda Cosmides and John Tooby, of the Center for Evolutionary Psychology at the University of California at Santa Barbara, and anthropologist Robert Kurzban, of the University of California at Los Angeles, wanted to test whether this was innate or whether it was just an artifact of how society groups individuals by race.
