The literature demonstrating sex differences in pain is sizable. Most explanations for these differences have focused on biologic mechanisms, and only a few studies have examined social learning. The purpose of this study was to examine the contribution of gender-role stereotypes to sex differences in pain. This study used experimental manipulation of gender-role expectations for men and women. One hundred twenty students participated in the cold pressor task. Before the pain task, participants were given 1 of 3 instructional sets: no expectation, 30-second performance expectation, or a 90-second performance expectation. Pain ratings, threshold, and tolerance were recorded. Significant sex differences in the "no expectation" condition for pain tolerance (t = 2.32, df = 38, P <.05) and post-cold pressor pain ratings (t = 2.6, df = 37, P <.05) were found. Women had briefer tolerance times and higher post-cold pressor ratings than men. When given gender-specific tolerance expectations, men and women did not differ in their pain tolerance, pain threshold, or pain ratings. This is the first empirical study to show that manipulation of expectations alters sex differences in laboratory pain.