Randomised Policy Trials

n the eighteenth century, the notion that new medical treatments should be evaluated through randomised trials began to gain acceptance. Despite criticism from doctors, who believed that their scientific expertise should be taken on faith, evidence-based medicine gradually gained adherents. Today, randomised trials are a necessary step in the licensing of drugs in Australia and throughout the developed world. By contrast, in the Australian policy sphere, robust evaluation of the effectiveness of particular programs remains rare. While policy evaluations have become more common over recent years, the efficacy of most of these evaluations remains questionable. Because randomised trials are so rare in Australia, the political rhetoric is usually substituted for hard evidence. Yet in the United States (US), where randomised trials are most common, researchers in education, health and welfare have convincingly demonstrated that they are the most effective way of testing whether policies achieve their stated goals. The discussion in this paper outlines why randomised trials tend to be superior to other forms of policy evaluation, and address six common objections to their use. The early evolution of randomised evaluation, first in medicine, and then in the social sciences is then discussed. This is followed by an analysis of several recent examples in which policy knowledge has been advanced through the use of randomised trials. The paper then concludes with some suggested areas in which evidence-based policy could be implemented in Australia.