Past research demonstrates that prejudice toward women and Black Americans often co-occur in individuals. The present studies examine factors related to accuracy in estimating the co-occurrence, or overlap, of prejudice toward women and Black Americans. Across two studies, criterion overlap percentages were computed using national datasets and separate participant samples estimated prejudice overlap. Results indicate that beliefs about the generalized nature of prejudice can improve accuracy by reducing faulty underestimation of the overlap in anti-Black racism and sexism. In addition to greater displayed accuracy in perceptions of prototypical perpetrators of prejudice (i.e., estimates of White men compared to White women), the present work suggests that accuracy is improved when estimating sexist attitudes from racist attitudes, rather than vice versa. Together, this work documents the accuracy of prejudice overlap perceptions, for the first time, and factors that facilitate accuracy (i.e., perpetrator prototypicality, known prejudicial attitude), with implications for intergroup dynamics research.