Assembles lessons about the technology of social experimentation by drawing on experience at RAND, notably the authors' experience in designing and managing the Health Insurance Study. The report is in three parts. The first part discusses when to conduct a social experiment and when not to. The second part discusses the management of an experiment; it emphasizes the importance of the traditional managerial skills, in spite of the exaggerated mystique that the literature has assigned to the managers of unusual experiments. The third part is a compilation of several practical "tips" for the prospective experimenter.
RAND Report
1980
Archibald, R, Newhouse, JP
http://www.rand.org/pubs/reports/R2479.html