Research on academic learning indicates that many students experience major difficulties with introductory statistics and methodology courses. We hypothesized that students' difficulties may depend in part on the fact that statistics tasks are commonly viewed as related to the threatening domain of math. In two field experiments which we carried out in the ecological context of introductory statistics courses, the representation of a statistics task administered to undergraduate students in Psychology (Study 1; n = 127) and Biology (Study 2; n = 166) was manipulated. In both studies we found that students who reported the least favorable prior attainments in math had their performance thwarted when the task was presented as related to the math domain, whereas no performance impairment was observed when the task was presented as related to the verbal/linguistic domain. Implications of the findings for the design and delivery of introductory statistics and research methodology courses are discussed.