Stereotype threat has been one of the most researched phenomena in the social psychology literature over the past one and a half decades. The number of citations of Claude Steele and Joshua Aronson’s seminal article speaks for itself: as of June 27, 2012 Google scholar counted 2990! Recently, economists have started to look at this phenomenon, emphasizing its importance for economic and social outcomes.
What exactly is stereotype threat and what are its consequences?
Stereotype threat is the well-documented phenomenon where an individual shows lower performance on tests that are diagnostic of a skill domain in which the social group he or she belongs to is negatively stereotyped. There is no consensus on the precise mechanism through which stereotype threat reduces performance, but roughly speaking the theory says it comes from the fact that the situation diverts attention away from executing the task to a concern about potentially fulfilling a negative stereotype.
To see the striking effects that stereotype threat seems to have let us look at some experimental evidence. The first and probably most prominent example concerns the intellectual test performance of African Americans. The treatment, that is, the condition under which the experimenters expected stereotype threat to affect performance, consisted simply of mentioning in the instructions that the test was diagnostic of intellectual ability. The hypothesis was that given the negative stereotypes about blacks’ intellectual abilities, under the treatment condition blacks, and not whites, would perform worse than compared to a situation in which no mention was made that the test was diagnostic of intelligence.
Indeed, Aronson and Steele found that the treated group performed significantly worse on the test than the black and white control group, and they found no difference in performance for whites in the treatment and control groups.
The second example is taken from an experiment testing stereotype threat effects on women’s math performance. The treatment in this experiment simply consisted of reminding women of their gender. The hypothesis was that reminding women of their membership to a group that is generally thought to perform worse than men in maths would generate stereotype threat and a decline in performance. Again, the hypothesis was confirmed by the experiment.
The third example comes from an experiment by economists Karla Hoff and Priyanka Pandey. They show that when young children in India were reminded of their caste membership, those from low castes performed worse on solving mazes than in a situation when they were not reminded of their caste membership. In India, individuals from low castes are stereotyped to have lower intellectual ability than upper caste members. Again, under the treatment condition the stereotype was confirmed, while there were no performance differences when caste was not made salient.
How may stereotype threat be linked to poverty?
Stereotype threat creates an obstacle for individuals that belong to social groups that already face the burden of historic disadvantage. This means that even when poor children, blacks or women are able to access careers or universities from which they have been historically excluded, they may not perform as well as members of the traditionally advantaged group. Even though the differences in performance do not come from initial differences in ability, stereotype threat may hence lead to self-confirming beliefs. This in turn can lead to persistent inequalities in performance across different groups, which eventually can translate into persistent inequalities in income.
Consider the following example: A black student takes a college entrance exam testing intellectual ability. Existing stereotypes about blacks say that their intellectual abilities are lower than that of other groups. Because of stereotype threat the black student actually performs worse on the test than, for instance, a white person with similar abilities. His or her chances of being admitted to a competitive college are lower, and hence his or her chances of finding a highly paid job after college are lowered as well. Additionally, the initial negative stereotype about blacks is confirmed (they perform worse than whites on average), and future cohorts of black applicants will suffer stereotype threat again. A vicious circle is created.
How could stereotype threat be avoided then? Is it sufficient to avoid clues that remind individuals of their stereotyped identity? Even though many stereotypes are so deeply entrenched in their subjects’ minds that no explicit cues are necessary to evoke them, experimental evidence suggests that interventions as simple as asking for the gender of test takers after the test as opposed to before would be sufficient to close gender gaps in test performance by as much as 33%. Such interventions are almost costless, but could indeed bring about substantial gains. In the long run, closing the performance gap may contribute to closing the outcome gaps and reduce inter-group inequalities.
 C. Steele, J. Aronson. Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology, 6(5):797–811, Nov. 1995.
 S. Spencer, C. Steele, and D. Quinn. Stereotype threat and women’s math performance. Journal of Experimental Social Psychology, 35:4–28, 1999.
 K. Hoff and P. Pandey. Discrimination, social identity, and durable inequalities. The American Economic Review, 96(2):206–211, May 2006.
 Danaher, K. & C. Crandall. Stereotype threat in applied settings re-examined. Journal of Applied Social Psychology, 38(6):1639–1655, 2008.