A Framework for Processing Value Judgments

Traditional decision-analytic practice emphasizes the distinction between probability assessments and value (or utility) judgments. Whereas techniques for elicitation and integration of subjective probabilities often can be submitted to empirical tests of validity, the fidelity of encoding value judgments has so far defied measurements. A unified approach to the treatment of the two types of judgments is presented; value judgments are interpreted as conditional probability statements. Such formulation leads to rational methodologies and procedures for solving the following tasks: 1) empirical validation and refinement of value judgments and 2) aggregating value judgments obtained from a panel of experts.