Causal Processes and Techniques in the Modelling of Policy Outcomes

Abstract

The statistical modelling of the causes of variations in policy outputs is still at an early and unsophisticated stage of development. Although it has already yielded much in the area — as much as older techniques of analysis — its future contribution will be far greater. But this will not happen until there exists a larger body of really careful, intelligent and scholarly research literature which is universally understood by a substantial number of research workers who have mastered the techniques of model-building with sufficient thoroughness to incorporate complex causal processes with precision. If this were a new area of physics, we should not have to wait long. But for bad reasons as well as good we are prejudiced against complex methodologies, and energetically promote new areas and methods before we have begun to reap a harvest from work based on the last set of new ideas. We may therefore have to wait. But there is no doubt that the preconditions will one day exist.