Considering the Case Against Experimental Evaluations of Social Innovations

The present author was asked to comment on the preceding article by Weiss and Rein (also see Weiss and Rein, 1969) because he has vigorously advocated a future experimenting society which would use hardheaded research designs to evaluate social innovations (Campbell, 1969a) whereas Weiss and Rein seem to reject such an approach. The main disagreement turns out to be with the packaging and labeling of their argument, rather than with its contents. Most of their illustrations of weakness in specific ameliorative programs are sensible. They do not oppose experimental designs lacking these weaknesses, although they doubt that such designs are possible for broad-aim programs. There is disagreement with their concept of the experimental method and the alternative methodology they propose. The latter seems to contain mutually incompatible elements, which are individually compatible with the experimental method and, indeed, require the experimental method for reducing equivocalities of inference. In the following, four of Weiss and Rein's important points are reformulated and interpreted in terms of alternative implementations of an experimental approach.1