Conditional Independence, the Mantel-Haenszel Test, and the Yates Correction

Abstract The Yates correction is often recommended as a means of improving the chi-squared approximation to the Pearson goodness-of-fit statistic in 2 × 2 tables. This article demonstrates that it is also needed for some conditional independence models in multiway tables containing two dichotomies. The assumptions of the Mantel-Haenszel test are discussed.

[1]  G. Barnard,et al.  On alleged gains in power from lower P-values. , 1989, Statistics in medicine.

[2]  W. G. Cochran Some Methods for Strengthening the Common χ 2 Tests , 1954 .

[3]  F. Yates Contingency Tables Involving Small Numbers and the χ2 Test , 1934 .

[4]  G. Upton A Comparison of Alternative Tests for the 2 Times 2 Comparative Trial , 1982 .

[5]  W. Haenszel,et al.  Statistical aspects of the analysis of data from retrospective studies of disease. , 1959, Journal of the National Cancer Institute.

[6]  J. V. Howard,et al.  The $2\times2$ table: a discussion from a Bayesian viewpoint , 1998 .

[7]  David E. Hapeman Statistical Analysis of Categorical Data , 2000, Technometrics.

[8]  Nathan Mantel,et al.  Chi-square tests with one degree of freedom , 1963 .

[9]  M. G. Haviland Yates's correction for continuity and the analysis of 2 x 2 contingency tables. , 1990, Statistics in medicine.

[10]  G. Upton Fisher's Exact Test , 1992 .

[11]  M. W. Birch The Detection of Partial Association, I: The 2 × 2 Case , 1964 .

[12]  W. Conover Some Reasons for Not Using the Yates Continuity Correction on 2×2 Contingency Tables , 1974 .

[13]  P. Bickel,et al.  Sex Bias in Graduate Admissions: Data from Berkeley , 1975, Science.

[14]  W. J. Conover,et al.  Some Reasons for Not Using the Yates Continuity Correction on 2 × 2 Contingency Tables: Rejoinder , 1974 .

[15]  F. Yates,et al.  Tests of Significance for 2 × 2 Contingency Tables , 1984 .

[16]  E. S. Pearson,et al.  The choice of statistical tests illustrated on the interpretation of data classed in a 2 X 2 table. , 1947, Biometrika.