The future of controlled randomized experiments: A briefing

In 1970-73, the Social Science Research Council convened meetings of scholars, civil servants, executives, service providers and others to summarize their experience in evaluating ameliorative social programs. A product of this effort was Social Experimentation (Riecken, Boruch, Campbell, Caplan, Glennan, Pratt, Rees, and Williams, 1974). The book, and more important, those who contributed to it influenced the technology of randomized field experiments and how contemporary projects and programs are evaluated. The task here is not to document to use of the book or the value of the work of those who contributed to it. The task here, thanks to Dr. Midge Smith, initiator of this collection of papers, is to speculate about the future of randomized controlled field tests for planning and evaluation. To forecast. Take a whack at predicting the course of events. This essay takes as the “initial conditions,” in the jargon of physics and engineering, the state of the art as represented in Social Experimentation. This is an idiosyncratic choice. But one must choose what is at one’s back in looking to a horizon. The forecast knowledge base for this essay includes the contributions made by colleagues in the United States and other countries between 1974 and 1994. All of their work cannot be cited. I apologize in advance for being unable to prepare the 100 page bibliography that might do justice to them. What follows are conjectures, for the next decade, about the use of controlled randomized experiments for planning and evaluating social programs. The predictions are put bluntly, this essay being conceived as a briefing rather than a disquisition.