Evaluation is necessary to produce stereotype threat performance effects

In the first direct test of the assumption that an evaluative source is necessary to produce stereotype threat effects, we combined a traditional stereotype threat manipulation with a scoring manipulation. Participants in a "pooled'' condition believed that their scores would be averaged across gender, whereas in a "segregated'' condition participants were led to believe that their scores would be segregated by gender. Thus we manipulated the potential for the experimenter to evaluate performance at the level of the group. Even though the gender stereotype was equally salient in the pooled and segregated threat conditions, performance was debilitated only in the segregated/threat condition, suggesting that evaluation is indeed necessary to produce stereotype threat performance effects.