When social scientists are hired as consultants, they face several ethical dilemmas. This article presents a logical argument for an ethical principle regarding the behavior of social scientists conducting policy evaluation research for a client. Given the isolated nature of evaluation research and the lack of scientific consensus within social science, the article argues that an ethical social scientist should avoid the most common types of consultant work in policy evaluations. Because a social scientist's responsibilities as a scientist may' conflict with his or her responsibilities as a citizen some ethical compromises are necessary and two such compromises are presented. Both of these compromises involve social scientists acting as teachers rather than as contracted researchers.