Interpretation of Conditional Probabilities in Probabilistic Interference Word Problems

Abstract

People's strategies on probabilistic inference word problems were investigated in an attempt to determine which of three theories explains their neglect of base rate information when estimating the probability of a hypothesis. These word problems present a base rate or prior probability (p(h)), some evidence (e) that typically conflicts with the prior expectation, and information on the reliability of the evidence, which is stated as p(e/h), the conditional probability of the evidence being seen if the hypothesis is true. The three theories are (a) subjects believe that the base rate is irrelevant; (b) they integrate base rate and evidence in a manner that underweighs the base rate; or (c) they misinterpret the reliability information p(e/h) as if it were p(h/e). Data using four distinct methods support the theory that subjects confuse the conditional probabilities. Keywords: Bayes theorem, Protocol analysis.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1990
Accession Number
ADA226907

Entities

People

  • Michelle A. Miller
  • Robert M. Hamm

Organizations

  • University of Colorado Boulder

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Cognitive Science
  • Databases
  • Education
  • Information Processing
  • Information Science
  • Mathematics
  • Medical Personnel
  • New York
  • Probability
  • Psychology
  • Reasoning
  • Reliability
  • Social Psychology
  • Social Sciences
  • Statistics
  • Students
  • Tape Recorders

Readers

  • Artificial Intelligence
  • Life Cycle Cost Analysis
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Information Retrieval