CONCERNING THE EFFECT OF SMALL CORRELATION ON CERTAIN LARGE SAMPLE TESTS AND CONFIDENCE INTERVALS FOR THE MEAN

Abstract

Most of the well known significance tests and confidence intervals for the population mean are based on the assumption of a random sample. The paper considers how the significance levels and confidence coefficients of a commonly used class of these tests and intervals are changed when the random sample requirement is violated and the number of observations is large. It is found that the introduction of even a slight amount of correlation can result in a substantial significance level and confidence coefficient change. Thus this class of tests and confidence intervals would seem to be of questionable practical value for large sets of observations. For two types of situations of practical interest, methods areutlined for obtaining large sample tests and confidence intervals for the mean which are not sensitive to the presence of correlation. These results are as efficient (asymptotically) as the tests and intervals they replace and are applicable to the general situation where the observations are not from the same population.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 17, 1949
Accession Number
AD0603840

Entities

People

  • John E. Walsh

Organizations

  • RAND Corporation

Tags

DTIC Thesaurus Topics

  • Coefficients
  • Convergence
  • Data Science
  • Distribution Functions
  • Inequalities
  • Information Science
  • Intervals
  • Lepidoptera
  • Notation
  • Observation
  • Probability
  • Statistical Analysis
  • Statistical Samples
  • Statistical Tests
  • Statistics
  • Theorems

Fields of Study

  • Mathematics

Readers

  • Calculus or Mathematical Analysis
  • Computational Modeling and Simulation