Can the activation of a prevalent stereotype alleging female math inferiority influence the math performance and task choice behavior of women? If so, what mediates each of these effects? In addition, what strategies can be used to reduce the impact of this stereotype on the performance of women? Three studies examined these questions by using techniques derived from stereotype threat (Steele, 1992), self-affirmation (Steele & Liu, 1983), misattribution (Schachter, 1964) and stigma-threat (Blascovich et al., 2001a) research. In Studies 1 and 2, collegiate women and men were (or were not) presented with a gender differences (or no gender differences) instructional set either prior to completing a math test or prior to selecting an upcoming task, respectively. Study 1 demonstrated that women performed more poorly on a math test after receiving the gender differences instructional set when compared to their male counterparts. However, no gender differences emerged when women and men received a gender fair instructional set. In addition, Study 1 revealed that the gender X instructional set interaction effect on performance was mediated by task confidence perceptions--although the confidence perceptions of men heavily influenced this effect. Study 2 found a trend that suggests that the instructional set manipulation may also have implications for participants' choice behavior. Whereas women appeared to be more likely to choose a math task over a proofreading task, when presented with a gender differences instructional set, women displayed the opposite choice pattern after receiving a gender fair instructional set. The trend amongst men suggested that they were more likely to choose a math task over a proofreading task irrespective of instructional set. Study 3 examined whether the performance deficits experienced by women could be reduced by employing either self-affirmation or misattribution processes. The results demonstrated that these deficits were alleviated when women were allowed to affirm the self prior to completing a math task. These findings are discussed in relation to stereotype threat theory and to potential educational interventions. Future directions for stereotype threat research are also discussed.