A nonasymptotic inference procedure for Cochran's Q test for the equality of matched proportions is described. An algorithm and FORTRAN 77 program are provided to compute Cochran's Q test statistic and the associated nonasymptotic probability value. The nonasymptotic method provides improvement over the usual asymptotic chi-squared analysis procedure whenever the effective number of subjects is small or the number of successes is small.
References
1.
CochranW. G. (1950) The comparison of percentages in matched samples. Biometrika, 37, 256–266.
2.
McNemarQ. (1947) Note on the sampling error of the difference between correlated proportions or percentages. Psychometrika, 12, 153–157.
3.
MielkeP. W.Jr.BerryK. J. (1995) Nonasymptotic inferences based on Cochran's Q test. Perceptual and Motor Skills, 81, 319–322.
4.
MyersJ. L.WellA. D. (1991) Research design & statistical analysis. New York: Harper Collins.
5.
PatilK. (1975) Cochran's Q test: exact distribution. Journal of the American Statistical Association, 70, 186–189.
6.
SiegelS.CastellanN. J. (1988) Nonparametric statistics for the behavioral sciences. (2nd ed.) New York: McGraw-Hill.
