This paper investigated stereotype threat effects involved in driving ability. In Experiment 1, female university students with an average of several years of driving experience completed a driving exercise on a simulator. Some women were told that the study was designed to assess "mental processes involved in driving" (control) whereas the other women were told that the study was to "investigate the reasons that men are better drivers than women" (stereotype threat). Although women in the two groups showed equivalent driving quality in terms of maintaining proper speed and road orientation, women in the stereotype threat condition were more than twice as likely to hit a pedestrian jaywalker who unexpectedly appeared near the end of the simulation (59% vs. 25%). Experiment 2 replicated the first study but added a concurrent demanding cognitive task for half of the participants. When participants were not asked to complete the demanding task while driving, results mirrored those obtained in Experiment 1; women in the stereotype threat condition were more than twice as likely to strike an unexpected pedestrian compared with women in the control condition. When drivers had to complete a concurrent cognitive task, however, performance in the control condition was negatively affected but was no worse in the stereotype threat condition. The fact that accidents were as likely in the control as in the stereotype threat condition under cognitive demand suggests that stereotype threat harms driving ability by reducing available cognitive capacity to react and adjust effectively to unexpected events.