Three experiments investigated the effects of exposure to sexism on women's task performance, self-esteem, and sense of belonging in an academic setting. In Experiment 1, female undergraduates about to participate in a study were told by another female student (a confederate of the experimenter) either that the male experimenter was sexist (stereotype threat) or no mention was made of the experimenter's attitudes towards women (control). Subsequently, the experimenter either provided negative or positive feedback regarding the students performance on a set of "practice problems" (similar to those used in the analytical section of the GRE). Performance on a later test showed that women who thought the experimenter was sexist performed more poorly than women in the control condition. Feedback had no effect on performance but did affect self-esteem; esteem was lower when students had received negative feedback. In Experiment 2, male and female students were exposed to the same sexism manipulation as in Experiment 1 but there was no manipulation of feedback. Although women again showed a tendency to perform more poorly when the experimenter was supposedly sexist, men actually performed better in the threat compared with the control condition. Self-esteem was again unaffected by the sexism manipulation, but women reported feeling less comfortable in the testing situation when sexism had been invoked. Experiment 3 varied the sex of the confederate so that male and female students in the threat conditions were informed by same-sex confederates that the experimenter was sexist (i.e., implying sexism against men for male students). Women again performed more poorly with a supposedly sexist experimenter, but men's performance was unaffected by this factor. Both men and women felt less comfortable in the situation if they thought the experimenter was exist, but self-esteem was unaffected by supposed experimenter sexism. In sum, these studies show that exposure to a sexist instructor can harm women's task performance and reduce their sense of comfort and belonging (but not necessarily their self-esteem). Although men also showed a lowered sense of belonging, their actual performance actually tended to improve under perceived sexism (i.e., they showed stereotype lift), perhaps because the gender stereotypes are suggestive of males' superior reasoning ability.