Lessons for Future Public Policy and Research

Social experimentation with random assignment is unique to the United States as an approach to public policy research. As far as I can determine, no other Western countries have conducted social experiments with a randomly assigned group. In the United States, we have now had sufficient experience with this form of policy research that it is appropriate to take a hard look at its conduct and its usefulness in the governmental process. This paper examines the income maintenance experiments and their implications for social policy and for future research. Each topic is introduced by a section on their lessons for the history of applied social science. To organize the material I have adopted the approach recommended by Richard E. Neustadt and Ernest R. May.1 They urge policy analysts to draw on the historical record to answer three questions: What is known? What is unclear? What is presumed? Before proceeding, it is necessary to add a comment about the point of view I bring to this analysis. As a former government official involved in the development of Nixon’s Family Assistance Plan, I have come to the conclusion that the Plan was a mistake. I believe the mistakes made in developing and advocating the Family Assistance Plan were a result of its heavy dependence on the idea of the negative income tax, which was riding high in the public policy research and analysis community at the time Nixon’s plan originated. As regards the second main topic considered in this paper -the research implications of the negative income tax experiments -I have a