We investigated whether risk-taking measures inadvertently focus on behaviors that are more normative for men, resulting in the overestimation of gender differences. Using a popular measure of risk-taking (Domain-Specific Risk-Taking) in Study 1 (N = 99), we found that conventionally used behaviors were more normative for men, while, overall, newly developed behaviors were not. In Studies 2 (N = 114) and 3 (N = 124), we demonstrate that differences in normativity are reflected in gender differences in self-reported risk-taking, which are dependent on the specific items used. Study 3 further demonstrates that conventional, masculine risk behaviors are perceived as more risky than newly generated, more feminine items, even when risks are matched. We conclude that there is confirmation bias in risk-taking measurement.