Although widely accepted in medicine and health services research, randomized controlled trials (RCTs) are often viewed with hostility by social scientists, who cite a variety of reasons as to why this approach to evaluation cannot be used to research social interventions. This article discusses the three central themes in these debates, which are those of science, ethics, and feasibility. The article uses three recent U.K. trials of social interventions (day care for preschool children, social support for disadvantaged families, and peer-led sex education for young people) to consider issues relating to the use of random allocation for social intervention evaluation and to suggest some practical strategies for the successful implementation of “social” RCTs. The article argues that the criteria of science, ethics, and feasibility can and should apply to social intervention trials in just the same way as they do to clinical trials.