This experiment tested whether performance on tasks under stereotype threat mirror the effects of arousal demonstrated in other domains. Specifically, it has been shown that arousal tends to decrease performance on difficult tasks, but increase performance on simple tasks. Male and female students were told that they would complete a series of math tests that have either shown gender differences (stereotype threat for women), or have not shown gender differences (control). The difficulty of the tests were varied so that half of the participants completed an easy set of math problems and the other half completed a difficult set of problems. Stereotype threat improved performance of women on the easy set of problems but harmed performance on the difficult set. Men were unaffected by the stereotype threat manipulation. These results are compatible with accounts of stereotype threat that point to a central role for arousal.