Research on the Technology of Inference and Decision
Abstract
Research on the technology of inference and decision is summarized. The first major research result, flat maxima in decision analysis, is summarized and its implications for social and psychological theory as well as decision technology are discussed. The major conclusions from this area of research is that the structuring of the decision problem may be much more important than the elicitation of specific parameter values. The second major topic is a review of multi-attribute utility measurement and an experiment is outlined which may circumvent the problem of evaluating utility elicitation methodologies without having an external standard for utility. Elicitation technology for probabilities constitutes the third research area. Extensive review of data collected in five critical experiments indicated that feedback may be more important than aggregation as a source of human conservatism in inference. The importance of this finding is stressed. The fourth and last area summarizes the final three of nine Technical Reports to stem from this contract. These concern Bayesian statistical analysis for comparing deterministic models, a critique of the Bayesian foundations of decision theory, and error analysis in probabilistic information processing systems.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 30, 1973
- Accession Number
- AD0771463
Entities
People
- Ward Edwards
Organizations
- University of Michigan