BINOMIAL ERROR MODELS IN MENTAL TESTING.

Abstract

A general model for the relation between fallible measurements and 'true' measurements is outlined. This model is appropriate in many scientific areas. It leads to the empirical Bayes problem of estimating the frequency distribution of true measurements for a population of measured objects. A special case important in mental test theory is considered. Practical results obtained by one estimation procedure are briefly outlined. (Author)

Document Details

Document Type
Technical Report
Publication Date
May 01, 1969
Accession Number
AD0688430

Entities

People

  • Frederic M. Lord

Organizations

  • Educational Testing Service

Tags

DTIC Thesaurus Topics

  • Binomials
  • Frequency
  • Measurement

Readers

  • Regression Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference