Is the demand-revealing process 1 merely an intriguing theoretical curiosity, or is it an important advance in the art of collective decision-making? Experimentation with the process is needed to answer this question. This paper presents the results of an experiment in which college fraternities were induced to make decisions in their weekly meetings by the demandrevealing process (hereafter DR). The results suggest that DR is a workable procedure, that it increases the efficiency of binary decisions by an average of about 2.25 percent of the net 'intensity of feelings' about the issues under consideration, and that some version of DR would be viewed as an improvement in collective decision-making by a substantial majority of those who have tried it.
A number of other researchers have experimented with a variety of collective decision procedures. Bohm (1972) solicited preferences for a public good under six different rules for the relationship between the value a person stated and the payment that would be required of the person. Somewhat surprisingly, Bohm found that the mean payments offered under the different rules were not significantly different. However, the small number of participants in the experiment made the lack of statistical significance somewhat less surprising.
Ferejohn, Forsythe and Noll (1979 -hereafter FFN) compared experimentally the procedure that the stations in the Public Broadcasting Service use to decide which programs to produce with Vernon Smith's 'auc-* This research grew out of Gordon Tullock's suggestion that experiments with the demandrevealing process were needed. It started as a joint project of the two of us. As it developed, my share of the effort increased to the point where he suggested that I should accept all author's responsibility. Helpful suggestions of Charles Plott and Subroto Roy are gratefully acknowledged. Marvin Ballard and Jane Smith provided invaluable research assistance.
Stephen Sheppard provided both computer assistance and substantial general counsel.