Judgments of Probability and Utility for Decision-Making
Abstract
The report describes research intended to develop procedures for eliciting judgments of probability and utility that could be employed efficiently in a decision theoretic analysis. Experiments on probability estimation investigated the reinforcing effects of a proper scoring rule upon probability estimates; the use of Bayesian procedures to revise probability estimates; and, the relative merits of probabilities and odds as response modes. Research on the decomposition of utility estimates showed that the model of a weighted linear average is relatively insensitive to nonadditive combination rules but highly sensitive to the nonlinearity of utility functions; when utilities are estimated it makes relatively little difference whether or not the judgments are decomposed; and, the procedures of decomposing utility estimates were feasible for use on a real world problem where water-quality engineers rather than college students served as subjects.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 30, 1971
- Accession Number
- AD0735139
Entities
People
- Cameron R. Peterson
Organizations
- University of Michigan