A version of multiattribute utility theory is evaluated as an aid to social decision making using an Australian case study. A multiattribute utility function is assessed over the descriptive, discrete decision alternatives of the decision problem rather than over the risky consequences of the attributes as in the standard approach. The evaluation is made first from the perspective of the decision analysts as a test of the feasibility of the method, and second, from the perspective of social decision makers as a test of the value of the analysis to them. Several difficulties arose in administering the procedure, and the decision makers were sceptical about the derived prescriptions. Nevertheless, the research findings appear to support the proposition that multiattribute utility theory should be viewed as more than merely a normative device for social decision making by experts.