Research on affective forecasting indicates that people regularly mispredict the emotional impact of negative events. We extended this work by demonstrating several forecasting errors regarding women's affective reactions to ambivalent sexism. In response to a survey about sexism against women, students at a university in the Central U.S. (N =188) overestimated the negative impact of hostile sexism, and underestimated the negative impact of benevolent sexism, relative to women's reports of their actual experiences. Moreover, people mispredicted both the intensity of women's initial affective reactions to, and the duration of women's recovery following, ambivalent sexism. The data supported a model in which inaccurate estimates of initial intensity fully accounted for people's inaccurate estimates of recovery duration following ambivalent sexism.