AbstractThis paper was originally presented under the title "Decision" at the Third Operational Research Conference of the Universities of Lancaster and Sussex held at Windermere in April 1968. In it, the question of what is meant by "improved decision-making" is examined and the ways in which operational research can and does contribute to this are discussed. In the course of this the "American School" of Decision Analysis is reviewed. The paper concludes with some suggestions on computational procedures. This paper was written as a vehicle for discussion. In order to stimulate this, the authors have been deliberately provocative and presented an extreme point of view. The paper which follows this is a condensed version of some of the discussion which took place at the Conference. The authors would welcome further contributions.