Introduction: This longitudinal study designed and tested the validity of a new measure of pro-social risk taking -- risks that individuals take in order to help others. Methods: The sample was racially and ethnically diverse adolescents in the rural Southeastern United States (N = 867; Mage = 12.82 years, 10-14 years at Wave 1; 50% Girls, 33% White non-Latinx, 27% Latinx, 20% Black, 20% Mixed/Other race/ethnicity). Adolescents completed self-report measures of the new prosocial risk-taking scale at baseline and one- and two-year follow-ups. Results: Confirmatory factor analysis demonstrated excellent model fit with a 6-item single factor score. Further, the scale demonstrated good test-retest reliability at one and two-year follow ups. The scale also demonstrated convergent validity, such that prosocial risk taking was positively correlated with prosocial tendencies, empathy, and sensation seeking, and negatively correlated with negative risk-taking behavior and risk tolerance. Finally, we found significant differences by race/ethnicity (but not by gender) in prosocial risk taking, which were not attributable to measurement invariance, and should be interpreted in the context of ongoing societal inequalities between youth. Conclusions: The new Prosocial Risk-Taking Scale yielded reliable scores in our sample. It may be used in future research to investigate individual differences in adolescents' prosocial risk taking, developmental change in prosocial risk taking, and the significance of prosocial risk taking for adolescents' emotional and social adaptation.