Policy Research in the Context of Diffuse Decision Making.

A distinctive characteristic of educational policies in the United States in the past fifteen years has been the mandate for evaluation that has accompanied them. Most major policy initiatives in education, as in health and social services, have been attended by formal systematic evaluation of the effects of the policy for its intended beneficiaries. Over all, the federal government has been spending hundreds of millions of dollars annually to learn how well human service policies are achieving the ends for which they were designed. The upsurge in evaluation activity and expenditures has a rational cast. The presumed purpose of all this analysis is to improve the effectiveness of policy. Evaluation, the rhetoric goes, will identify the programs and policies that are working well so that they can be expanded, and locate the programs and policies that are working poorly so that they can be terminated or modified. Evaluations that analyze the effects of component strategies of intervention-that indicate which components of policies are successful for which types of clientele under which conditions-will provide the basis for modifying policies and attuning them to the needs and life conditions of the participants. The enterprise, in short, is meant to use the methods and techniques of social science in the service of rational allocation of resources and the improvement of welfare policy. American social scientists by the thousands have been attracted to