Allport's Legacy and the Situational Press of Stereotypes

Journal of Social Issues, Fall, 1999 by David M. Marx, Joseph L. Brown, Claude M. Steele

Is Stereotype Threat Mediated by Negative Expectancies?

As we have noted, the legacy most commonly associated with Allport's (1954) treatment of the targets of prejudice is that those who are negatively stereotyped hold corresponding negative expectations about themselves. Could it be, then, that the effects of the stereotype threat manipulations in the above experiments were in some way dependent on the negatively stereotyped groups--African Americans and women--holding the negative expectancies about their performance that are implicit in the negative stereotypes about their group?

Further research suggests this is not the case. Aronson and his colleagues (Aronson et al., 1999) devised a way of putting White males under stereotype threat while taking a difficult math exam. In the critical stereotype threat condition, the researchers told White, math-identified males that they would be taking a test on which Asians generally did better than Whites. Here, then, is a group that is not chronically exposed to negative stereotypes about its math ability, but in this particular situation, is under the stereotype threat of being seen negatively through the lens of a positive stereotype about another group. In this condition, White males significantly underperformed in comparison to White males not under stereotype threat. This finding makes the point that the situational threat of negative stereotypes is sufficient to impair test performance even among groups that are not chronically stereotyped negatively in the domain. Internalization of negative group stereotypes as low self-expectancies is apparently not a necessary component of stereotype threat effects.

But if chronically low expectancies are not critical to stereotype threat effects, perhaps the effect of stereotype threat manipulations in these experiments stems from their acute lowering of students' performance expectations in the specific experimental session. However, there is little evidence that stereotype-threatened students are withdrawing effort, as would be indicated if the stereotype threat conditions lowered performance expectancies. C. M. Steele and Aronson (1995) observed stereotype-threatened Black students taking a difficult verbal test on the computer and found that these students appeared to be working too hard, rather than too little, on the items, rereading the questions, rechecking their answers, spending more time presumably to ensure the accuracy of their work. Also, stereotype-threatened students typically attempted just as many problems as nonthreatened students did, only with less accuracy.

This pattern of results suggests, then, that stereotype threat occurs not because of reduced self-expectancies, but because the salience of the stereotype fosters a mistrust of the situation. Stereotype-threatened individuals cannot trust that interpretations of their performance will be free of the taint of negative stereotypes.

What the foregoing research has shown is that when a negative group stereotype is salient as a relevant interpretation of a person's performance then that performance is impaired. If this threat is made less salient, performance improves. The results thus far highlight the situational, and malleable, nature of stereotype threat. The fact that the most affected students are those who have high levels of confidence and high performance expectations casts doubt on the notion that chronic negative expectancies also are critical to these effects. However, these studies have not definitively ruled out the possibility that stereotype-relevant conditions prime lower expectancies. A number of recent theorists and researchers (e.g., Howard & Hammond, 1985; Stangor, Carr, & Kiang, 1998; S. Steele, 1990) have argued that lower self-confidence and performance expectancies are triggered by situationally relevant negative stereotypes.