Summary
A rigorous study is carried out for the comparison of the exact Bartlett distribution for the test of the homogeneity of variances for unequal sample sizes with that of equal sample size. A detailed estimate is derived to the correction term of the relevant distributions. LetM 1=−Σν j ln(S 2 j /S 2), wherek denotes the number of (normal) populations and ν j denotes the degree of freedom of thej-th sample. LetM 2 be the corresponding variable when allv j =v. In general, letv=minv j and δ=max(v j -v)/v j . We then obtain a bound\(|P [M^1 /C_1 \geqslant x] - P [M^2 /C_2 \geqslant x]| \leqslant \bar \varepsilon (\delta ,\underset{\raise0.3em\hbox{$\smash{\scriptscriptstyle-}$}}{v} ,k)\), uniformly inx, whereC 1=C 1(k,v j ),C 2=C 2(k,v) are suitable scaling factors. For sufficiently small\(\bar \varepsilon \), the critical values for unequal sample sizes may be then taken to beM 1 C =(C 1/C 2)M 2 C , where the equal sample sizeM 2 C values are well known. A rate of decrease of\(\bar \varepsilon (\delta ,\underset{\raise0.3em\hbox{$\smash{\scriptscriptstyle-}$}}{v} ,k)\) for δ→0 and δ→0 is also given. An example is given to illustrate the practical usefulness of our estimate.
Similar content being viewed by others
References
Bartlett, M.S.: Properties of Sufficiency and Statistical Tests. Proc. Roy. Soc. London, Ser. A.160, 1937, 268–282.
Box, G.E.P.: A General Distribution Theory for a Class of Likelihood Criteria. Biometrika36, 1949, 317–346.
—: Nonnormality and Tests on Variances. Biometrika40, 1953, 318–335.
Box, G.E.P., andS.L. Andersen: Permutation Theory in the Derivation of Robust Criteria and the Study of the Departures from Assumption. J. Roy. Stat. Soc., Ser. B17, 1955, 1–26.
Brown, G.W.: On the Power of theL 1 Test for Equality of Several Variances. Ann. Math. Stat.10, 1939, 119–128.
Chao, M.T., andR.E. Glaser: The Exact Distribution of Bartlett's Test Statistic for Homogeneity of Variances with Unequal Sample Sizes. J. Amer. Stat. Assoc.73, 1978, 422–426.
Dyer, D.D., andJ.P. Keating: On the Determination of Critical Values for Bartlett's Test.75, 1980, 313–319.
Gartside, P.S.: A Study of Methods for Comparing Several Variances. J. Amer. Stat. Assoc.67, 1972, 342–346.
Glaser, R.E.: Exact Critical Values for Bartlett's Test for Homogeneity of Variances. J. Amer. Stat. Assoc.71, 1976, 488–490.
—: A Characteristic of Bartlett's Statistic Involving Incomplete Beta Functions. Biometrika67, 1980, 53–58.
Gradshteyn, I.S., andI.M. Ryzhik: Tables of Integrals, Series and Products. New York 1972.
Harsaae, E.: On the Computation and Use of a Table of Percentage Points of Bartlett's M. Biometrika56, 1969, 273–281.
Hartley, H.O.: Testing the Homogeneity of a Set of Variances. Biometrika31, 1940, 249–255.
Keselman, J.H., P.A. Games, andJ.J. Clinch: Tests for Homogeneity of Variance. Commun. Stat. B8, 1979, 113–129.
Nandi, S.B.: On the Exact Distribution of a Normalized Ratio of the Weighted Geometric Mean to the Unweighted Arithmetic Mean in Samples from Gamma Distribution. J. Amer. Stat. Assoc.75, 1980, 217–220.
Pearson, E.S., andH.O. Hartley: Biometrika Tables for Statisticians. Vol. I, 3rd Ed. Cambridge 1966.
Pitman, E.J.G.: Tests of Hypotheses Concerning Location and Scale Parameters. Biometrika31, 1939, 200–215.
Thompson, C.M., andM. Merrington: Tables for Testing the Homogeneity of a Set of Estimated Variances. Biometrika33, 1946, 296–301.
Wilk, M.B., andS.S. Shapiro: The Joint Assessment of Normality of Several Independent Samples. Technometrics10, 1968, 825–839.
Author information
Authors and Affiliations
Additional information
Work supported by the “Chief Research and Development Award” under CRAD No: 3610-637.
Rights and permissions
About this article
Cite this article
Manoukian, E.B. Departure of Bartlett's distribution for the homogeneity of variances for unequal sample sizes from that of equal sample sizes. Metrika 30, 179–194 (1983). https://doi.org/10.1007/BF02056923
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/BF02056923