An examination of stereotype threat in a motivational context
This study examined whether stereotype threat might be less likely in situations where there are incentives for strong performance. In addition, the study was designed to determine if stereotype threat might also affect performance on a test irrelevant to the domain in question. To address these issues, Black and White undergraduates were told that the study examined performance on two different types of tests, and that high performers across both tests would receive a monetary reward. Half the students completed a racial identity scale and a demographic form before taking the tests and were told that tests were good indicators of intelligence and ability (stereotype threat for Blacks regarding cognitive ability). The other half of students were told that they were completing a personality and problem solving test, and that their racial and demographic information would be collected at the end of the study (control). The order of completing the cognitive ability and personality tests was also manipulated. Analysis of the performance on the cognitive ability test showed that Black students performed less well on the test than did Whites, but there were no effects of the threat manipulation. Whites who identified with the domain performed better on the cognitive ability test than those who did not, but there was no relationship between domain identity and performance for Blacks. Different effects emerged from analyses focusing on racial identity, depending on when racial identity was assessed. Blacks scoring high in racial identity before the test performed better on the cognitive ability test than those who were low in racial identity. In contrast, Blacks high in identity on a measure completed after the test performed worse than those Blacks low in racial identity. It appears that a strong racial identity aids in performance, but good performance can ironically decrease racial identity. Consistent with this interpretation is the finding that Black students who showed the biggest decrease in racial identity from the beginning to the end of the experiment performed best on the test. Thus, effective testing performance might result in (or even require) disidentification over the course of testing. On the personality test, individuals performed better in the threat compared with the control condition, and this effect was stronger for Black than for White students. Overall, these findings point to a relatively small role of stereotype threat when external incentives for performance are established. In such conditions, performance depends on both domain identity and racial identity. In addition, this study shows that stereotype threat can actually increase performance on tasks unrelated to the stereotype.