Reconciling Incoherent Judgments (RIJ) - Toward Principles of Personal Rationality.

Abstract

Decision analysis involves constructing models which force logical coherence between a subject's judgments, e.g., between his choice of action and probabilities and utilities. However, it does not specify how he should reconcile any incoherent judgments. There are indefinitely many ways they can be adjusted to be coherent systems of judgment. The authors discuss two approaches for identifying one ideal set of reconciled judgments for a subject, given some or, in the limit, all potential incoherent readings. They both call for higher order judgments bearing on the precision of the subject's original readings. One is a straightforward extension of Bayesian updating with the readings serving as data to update a prior. The other involes minimizing adjustments, taking into account the stability of the readings as probabilistically measured. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1978
Accession Number
ADA059639

Entities

People

  • Dennis V. Lindley
  • Rex V. Brown

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Applied Psychology
  • Biological Sciences
  • Decision Theory
  • Engineering
  • Fuzzy Sets
  • Human Factors Engineering
  • Military Research
  • Operations Research
  • Probability
  • Probability Distributions
  • Psychology
  • Random Variables
  • Reasoning
  • Statistics
  • Students
  • Surveys
  • Theorems

Readers

  • Approximation Theory.
  • Strategic Security Studies
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference