We develop an information-based intervention against sexual harassment and test it in a randomized control trial across small groups of military recruits in the boot camp of the Norwegian military. The intervention seeks to bridge two knowledge gaps with implications for sexual harassment prevalence. We tell some recruits about their peers’ beliefs that “telling sexualized jokes can be labeled sexual harassment” and about women soldiers’ equal performance on military skill tests. This treatment gives lasting improvements in knowledge about what sexual harassment is and about women’s job performance. The impact on sexual harassment prevalence is directionally negative but statistically insignificant. We discuss measurement error and use survey responses about a harassment scenario to argue that the intervention likely affected behavior. Our study provides the first field experiment to evaluate whether a prevention method against sexual harassment reduces prevalence in a work setting. We use insights from our research process to identify methodological pitfalls and provide guidance for future field experiments in this area.