ON THE 'JUST NOTICEABLE DEPARTURE' OF DATA FROM THEORY,

Abstract

A lower bound on the noncentrality parameter of the chi-square test of goodness of fit is derived. The bound is a function of the average of absolute discrepancies between theoretical and true probabilities. This bound provides a method that estimates the correspondence between theory and data, and that nullifies, in part, Grant's strictures against traditional goodness-of-fit tests. The method also serves as a guide to planning an empirical evaluation of a quantitative theoretical model, when a chi-square test of goodness of fit is contemplated. (Author)

Document Details

Document Type
Technical Report
Publication Date
Jul 15, 1967
Accession Number
AD0657934

Entities

People

  • John Van Laer

Organizations

  • University of Washington

Tags

DTIC Thesaurus Topics

  • Chi Square Test
  • Cooperation
  • Data Science
  • Goodness Of Fit Tests
  • Information Science
  • Mathematics
  • New York
  • Pathologic Constriction
  • Probability
  • Test And Evaluation

Fields of Study

  • Mathematics

Readers

  • Regression Analysis.