Group Differences in IQ: Stereo Type Threat

STEREO TYPE THREAT

In the previous section, we asked whether blacks’ and whites’ IQ scores would be the same if we could match their environments. If the answer is yes, then this obviously points toward an environmental explanation of the race difference. But to tackle this question in a thorough way, it may not be enough to match factors like parental educa-tion, income, and occupational level. Even if we succeed in matching for these aspects of life, black and white children still grow up in different environments. This is because, after all, black children grow up knowing they are black and knowing a lot about what life paths are easily open to them and what life paths are likely. White children corre-spondingly grow up knowing they are white, and they too have a sense of what life paths are open or likely. Moreover, each group, because of the color of their skin, is treated differently by the people in their social environment. In these ways, their envi-ronments and experiences are not matched—even if the parents have similar jobs and similar income levels, and even if the children have similar educational experiences.

Do these social experiences matter for intelligence scores? As one indication that they do, consider studies of stereotype threat, a term used to describe the negative impact that social stereotypes, once activated, can have on task performance. Here’s an example: Imagine an African American is taking an intelligence test. She might well become anxious because she believes this is a test on which she is expected to do poorly. This anxiety might then be compounded by the thought that her poor performance will only serve to confirm others’ prejudices. These feelings, of course, could then easily erode performance by making it more difficult for her to pay atten-tion and do her best work. Moreover, given the discouraging thought that poor performance is inevitable, she might well decide not to expend enormous effort—if she’s likely to do poorly, why struggle against the tide?

Evidence for these effects comes from various studies, including some in which two groups of African Americans are given exactly the same test. One group is told, at the start, that the test is designed to assess their intelligence; the other group is led to believe that the test is simply composed of challenges and is not designed to assess them in any way. The first group, for which the instructions trigger stereotype threat, does markedly worse (C. Steele, 1998; C. Steele & Aronson, 1995).

Related results have been shown in many other circumstances and have been demon-strated with children as well as adults. Similar data have also been reported for groups other than African Americans, and in fact stereotype threat is plainly relevant to our pre-vious discussion of comparisons between men and women (Blascovich, Spencer, Quinn,Steele, 2001; Cheryan & Bodenhausen, 2000). For example (and as we mentioned earlier), merely reminding test takers of their gender just before they take a math test seems to encourage women to think about the stereotype that women cannot do math, and this seems to undermine their test performance. This is because thoughts about the stereotype increase the women’s anxiety about the test, cut into the likelihood that they’ll work as hard as they can, and make it less likely that they’ll persevere if the test grows frustrating (Ambady, Shih, Kim, & Pittinsky, 2001, Figure 11.25). A different study had students read an essay that argued that gender differences in math performance have genetic causes; women who read this essay then performed more poorly on a math test than did women who read essays on other topics (Dar-Nimrod & Heine, 2006). Presumably, the essay on genetic causes was demoralizing to the women and made them more vulnerable to stereotype threat—and therefore undermined their test performance.

Conversely, some interventions can improve performance, presumably by dimin-ishing the anxiety and low self-expectations associated with stereotype threat. In one study, middle-school students were asked to write brief essays—just a few sentences—about things they valued. The participants were given a list of possible values to choose from: “athletic ability, being good at art, being smart, creativity” and so on (G. Cohen, Garcia, Apfel, & Master, 2006; Cohen, Garcia, Purdie-Vaughns, Apfel, & Brzustoski, 2009). This brief exercise was then repeated periodically during the school year; and this was enough to shift the students’ perspective, getting them to focus on things they valued rather than school-based anxieties. In fact, the brief intervention improved the school grades of African American seventh-graders by a striking 40%, markedly reducing the difference between white students’ and black students’ grades. Remarkably, effects of the intervention were still detectable in a follow-up study with the same students, two years later.

These results draw our attention back to our earlier comments about what intelligence is—or, more broadly, what it is that “intellectual tasks” require. One requirement, of course, is a set of cognitive skills and capacities (e.g., mental speed, or working memory). A different requirement is the proper attitude toward testing—and the wrong attitude (anxiety about failing, fear of confirming other’s negative expectations) can plainly undermine performance. This is why performance levels can be changed merely by priming people to think of themselves as members of a certain group—whether that group is women, Asians, or African Americans. In this way, social pressures and prejudice can powerfully shape each person’s performance—and can, in particular, contribute to the differences between IQ scores for whites and blacks.