Chapter 11. Categorical Data and Chi-Square Tests

  1. Michael R. Chernick1,
  2. Robert H. Friis2

Published Online: 11 APR 2003

DOI: 10.1002/0471458716.ch11

Book Title

Introductory Biostatistics for the Health Sciences: Modern Applications Including Bootstrap

Introductory Biostatistics for the Health Sciences: Modern Applications Including Bootstrap

Additional Information

How to Cite

Chernick, M. R. and Friis, R. H. (2003) Categorical Data and Chi-Square Tests, in Introductory Biostatistics for the Health Sciences: Modern Applications Including Bootstrap, John Wiley & Sons, Inc., Hoboken, NJ, USA. doi: 10.1002/0471458716.ch11

Author Information

  1. 1

    Novo Nordisk Pharmaceuticals, Inc., Princeton, New Jersey, USA

  2. 2

    California State University, Long Beach, California, USA

Publication History

  1. Published Online: 11 APR 2003
  2. Published Print: 21 FEB 2003

ISBN Information

Print ISBN: 9780471411376

Online ISBN: 9780471458715

SEARCH

CHAPTER TOOLS

Get PDF (377K)

Keywords:

  • chi-square;
  • chi-square distribution;
  • independence;
  • homogeneity;
  • contingency table;
  • Simpson's paradox;
  • McNemar's test;
  • relative risk;
  • odds ratio;
  • goodness-of-fit test

Summary

  • The chi-square test is one of the most commonly cited tests in the biomedical literature.

  • The chi-square test may be used for several specific applications:

  • The formula for many test statistics with approximate chi-square distributions

  • Testing independence between two variables

  • Contingency table or cross-tabulation of two variables

  • Chi-square test for homogeneity is used in empirical investigations when…

  • Chi-square tests also may be considered tests of proportion.

  • In Simpson's paradox we see a particular association in each table; when we combine…

  • McNemar's test is used for categorical data that are correlated.

  • The concepts of relative risk and odds ratios are derived from epidemiologic studies.

  • General caveats regarding use of the chi-square test

Get PDF (377K)