ABSTRACT
Evaluation is frequently regarded as primarily a research procedure for testing intervention impact. Impact assessment is important, particularly for higher level decision–makers in national bodies. The problem is, of course, that past counterfactual states can never be known with certainty. In public policy and public administration, classic experimentation implies a two–stage approach to decision making. Prior to permanent enactment, the program ought to be provisionally tried out on a small scale under full experimental conditions. The Hawthorne effect is related to self-fulfilling prophesies. A selfnitely fulfilling prophesy is a false definition of a situation evoking new behavior, which makes the originally false conception come true. The demands of randomized experimentation collide directly with the way public sector decisions are made. The requirements of radical experimentation that interventions be tried out provisionally with random or matched groups prior to full inauguration cannot be met in the real world of public sector programming.
