Representation is not sufficient for selecting gender diversity

One strategy for promoting female leaders in science and technology professions is to appoint more women to the committees that select leaders. Unfortunately, evidence from other settings, such as committees for selecting judges or professors, suggests this approach does not work. We use a natural experiment to test the idea that organizational norms supporting gender diversity are necessary for representation on "selectorates" to promote gender diversity among leaders in science and technology. Our empirical setting is the standard-setting organization that develops key protocols for Internet hardware and software. We find that when more women are randomly selected for the committee that appoints the organization's leaders, the committee appoints more female leaders, but only after a set of interventions meant to increase members' awareness of the benefits of gender diversity.