Toward an Improved Methodology to Construct and Reconcile Decision Analytic Preference Judgments

Psychologists and behavioral economists have documented a variety of judgmental flaws that people make when they face novel decision situations. Similar flaws arise when decision analysts work with decision makers to assess their preferences and trade-offs, because the methods the analyst uses are often unfamiliar to the decision makers. In this paper we describe a process designed to mitigate the occurrence of such biases; it brings together three steps. In training, the decision maker is first given values to apply in judgment tasks unrelated to the decision at hand, providing an introduction to thinking deliberately and quantitatively about preferences. In practice, the learned tasks are then applied to a familiar decision, with the goal of developing the next incremental level of expertise in using the methods. Finally, in application, the more deliberative style of thinking is used to address the problem of interest. In an environmental resource setting with two oyster habitat managers, we test the procedure by attempting to mitigate the prominence effect that has been reported in the behavioral research literature. The resulting preference weights appear to be free of the prominence effect, providing initial steps toward operationalizing the “building code” for preferences introduced by Payne et al. [Payne JW, Bettman JR, Schkade DA (1999) Measuring constructed preferences: Towards a building code. J. Risk Uncertainty 19(1–3):243–270].

Duke Scholars

Author Robert T. Clemen Fuqua School of Business