Five studies tested the role of thought suppression under stereotype threat. Suppressing stereotypical thoughts may be important in coping with stereotype threat, but the act of suppressing unwanted thoughts such as stereotypes can reduce cognitive resources that are themselves necessary for optimal performance. In Study 1, college students who identified highly with math (N = 50, 25 male, 25 female) completed an easy math test under normal testing conditions (control) or while having to monitor a second task while taking the test (under a cognitive load). Results showed a slight tendency for women to do worse under cognitive load, suggesting that they were simultaneously coping both with a cognitive load and other intrusive thoughts. Consistent with this interpretation, women in the cognitive load condition were fastest to recognize stereotype-relevant words (e.g., illogical, irrational, worried) on a subsequent response time task. Study 2 focused on male and female students (N = 71, 35 male, 36 female) taking a challenging math test that was expected to instigate stereotype threat in women. Students took a math test and completed a word recognition response time task in that or reverse order. For women who completed the word recognition task first, response times to stereotypical words were slower. Women who had already completed the math test showed faster judgments of stereotypical words. These results suggest that women were successful in suppressing math-related stereotypes while preparing for the math test, but these thoughts were highly accessible after the test was completed. Study 3 involved students (N = 54) taking the same easy math test as in Study 1 under cognitive load (stereotype threat for women), except some of the students were told that the test was gender-neutral (control). Women who were told that the test was unbiased performed better than women who were not given this information. These women also showed no evidence that stereotypic thoughts were accessible in the response time task given at the end of the study whereas women in the stereotype threat condition responded more quickly in identifying stereotypic words. Studies 4 (N = 75 men and women) and 5 (N = 90 women) focused on strategies for students to avoid the costs of stereotype suppression under stereotype threat. Study 4 involved telling a subset of students that if they found themselves worrying about their performance on the math test, they should think instead about an identity that is personally important to them. Female students who had been given this instruction performed equally with men on the math test and better than women who were not given this instruction and showed no evidence that stereotypes had been activated based on their performance in the response time task. Study 5 showed similar results with a neutral thought replacement instruction that did not involve a valued personal identity. These results show that any strategy that distracts for stereotypical thoughts can enhance performance by reducing the need to suppress stereotypic concerns. In general, thought suppression is a potentially important mediator of stereotype threat, leading people to underperform when they exert effort to avoid thoughts of negative stereotypes.