Previous research has illuminated an important connection between stereotypes and the performance of those targeted by a stereotype. This body of work suggests that even implicit (i.e., nonconscious and unintended) math-gender stereotyping is related to poor math performance among women. Our longitudinal study sought to measure students' math-gender stereotyping during a college math course and examine the relationship between changes in implicit stereotyping and course performance. Results showed that, for both male and female students, stereotypes increased during the course. Importantly, there was a significant interaction between gender and changes in implicit stereotyping when predicting course performance. Female students showed a negative relationship between changes in implicit stereotypes and course performance, while male students showed no relationship between changes in implicit stereotyping and course performance. This suggests that only for women, who are stereotyped as poor math performers, did the observed increases in stereotyping over time predict poorer math performance.