Decision Analysis: Engineering Science or Clinical Art

Abstract : This paper examines two philosophical poles concerning the structuring of decision-analytic models. The engineering science approach uses complex, engineering-like models to link the decision maker's alternatives to his or her value structure; a computer then calculates the decision-analytic answer. The clinical art approach develops a simple model that structures the decision maker's thoughts concerning a decision in such a way that the critical issues in choosing one alternative over another are readily apparent. Three exemplary applications of each philosophical pole are used to demonstrate that at the extremes, the work of practitioners can be categorized as we describe. A discussion of the two practical extremes of decision analysis is then presented. In this discussion, dichotomous characteristics are assigned to the decision maker, the decision, and the analyst's constraints to portray conditions under which either the engineering science approach or the clinical approach would be more effective in producing appropriate insight and understanding. Then the implications of each approach are described in terms of the tools and procedure the decision analyst uses. Finally, the selection of the approach the decision analyst might use in any given application is graphically portrayed and discussed in terms of the points made in this paper. (Author)