JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. American Statistical Association is collaborating with JSTOR to digitize, preserve and extend access to Journal of the American Statistical Association. The evaluation of social programs is typically hampered by serious methodological problems. Adopting the approach of Bayesian simultaneous estimation proposed by Lindley and Smith (1972) and operationally developed by Jackson, Novick, and Thayer (1971), Novick, Lewis, and Jackson (1973), and Lewis, Wang, and Novick (1975), we demonstrate how some of these problems can be avoided or solved. To accomplish this, we extend the operational method to marginal m-group mean analysis for the unbalanced case. Data from the Emergency School Aid Act (ESAA) evaluation are used for illustration. Bayesian m-group analysis techniques are applied to data gathered from 17 school districts to evaluate the effectiveness of each district's compensatory education program. Ozenne is Project Officer, Office of Planning, Budget, and Evaluation, U.S. Office of Education. Research was supported, in part, by the U.S. Office of Education contracts OEC-0-73-0831 and OEC-0-73-6336 to the System Development Corporation. Points of view or opinions do not necessarily represent official Office of Education position or policy or that of SDC. This system's development has benefited from discussion of the problems in the evaluation of compensatory education by the Research Design Panel of ESAA, notably with William Coffman, Seymour Feshbach, Chester Harris, and particularly David Wiley. We also thank Michael J. Wargo of the U.S. Office of Education and John Coulson of SDC for support of this research.
