# Aversion to Epistemic Uncertainties in Rational Decision Making: Effects on Engineering Risk Management

*by*M. Elisabeth Paté-Cornell, Stanford Univ, United States,

Paul S. Fischbeck, Stanford Univ, United States,

**Abstract:**When faced with epistemic uncertainties in a decision problem, subjective expected-utility theory (SEU) implies that the only relevant characteristic of these uncertainties is the mean of the outcome probability distribution. Behavioral decision theories, however, have repeatedly shown that people are averse to epistemic uncertainties (or ambiguity, i.e., lack of fundamental knowledge). For example, in the regulatory context, one may be consistently willing to trade an ambiguous probability of system failure with a given mean for a well-understood failure probability with a value higher than this mean. In this paper, a normative theory for incorporating ambiguity attitudes into rational decision making is briefly presented and its implications for engineering risk management are discussed. A key result of our theory is the determination of an individual's epistemic risk premium, i.e., the amount of safety or money that an individual is willing to forgo in order to face better-defined failure probabilities. For engineering systems for which the objective is to meet a safety threshold (i.e., a maximum acceptable probability of failure), this premium can be represented as an adjustment of the threshold value. For example, for probability-based building codes, SEU implies that the mean of the failure probability distribution must meet the safety threshold.

**Subject Headings:**Risk management | Probability | Failure analysis | Uncertainty principles | Safety | Probability distribution | Decision making | Systems engineering

**Services:**Buy this book/Buy this article

Return to search