One-way ANOVA with Unequal Variances

We consider the one-way ANOVA problem of testing the equality of several normal means when the variances are not assumed to be equal. This is a generalization of the Behrens-Fisher problem, but even in this special case there is no exact test and the actual size of any test depends on the values of the nuisance parameters. Therefore, controlling the actual size of the test is of main concern. In this article, we first consider a test using the concept of generalized p-value. Extensive simulation studies show that the actual size of this test does not exceed the nominal level, for practically all values of the nuisance parameters, but the test is not too conservative either, in the sense that the actual size of the test can be very close to the nominal level for some values of the nuisance parameters. We then use this test to propose a simple F-test, which has similar properties but avoids the computations associated with generalized p-values. Because of its simplicity, both conceptually as well as computationally, this F-test may be more useful in practice, since one-way ANOVA is widely used by practitioners who may not be familiar with the generalized p-value and its computational aspects.

[1]  Xiao-Li Meng,et al.  Posterior Predictive $p$-Values , 1994 .

[2]  Edward J. Dudewicz,et al.  Exact solutions to the Behrens–Fisher Problem: Asymptotically optimal and finite sample efficient choice among , 2007 .

[3]  Joachim Hartung,et al.  Small sample properties of tests on homogeneity in one—way Anova and Meta—analysis , 2001 .

[4]  A New Generalized p-Value and Its Upper Bound for ANOVA Under Unequal Error Variances , 2008 .

[5]  W. R. Rice,et al.  One-way analysis of variance with unequal variances. , 1989, Proceedings of the National Academy of Sciences of the United States of America.

[6]  H. Jeffreys,et al.  Theory of probability , 1896 .

[7]  Morton B. Brown,et al.  The Small Sample Behavior of Some Statistics Which Test the Equality of Several Means , 1974 .

[8]  Kam-Wah Tsui,et al.  Generalized p-Values in Significance Testing of Hypotheses in the Presence of Nuisance Parameters , 1989 .

[9]  M. Vangel.  A numerical approach to the Behrens-Fisher problem , 2005 .

[10]  W. G. Cochran.  Problems arising in the analysis of a series of similar experiments , 1937 .

[11]  Sam Weerahandi,et al.  Size performance of some tests in one-way anova , 1998 .

[12]  Welch Bl.  THE GENERALIZATION OF ‘STUDENT'S’ PROBLEM WHEN SEVERAL DIFFERENT POPULATION VARLANCES ARE INVOLVED , 1947 .

[13]  Hubert J. Chen,et al.  Single-stage analysis of variance under heteroscedasticity , 1998 .

[14]  J. Linnik.  Statistical Problems With Nuisance Parameters , 2008 .

[15]  K. Krishnamoorthy,et al.  A parametric bootstrap approach for ANOVA with unequal variances: Fixed and random models , 2007, Comput. Stat. Data Anal..

[16]  Sam Weerahandi.  ANOVA under Unequal Error Variances , 1995 .

[18]  B. L. Welch.  ON THE COMPARISON OF SEVERAL MEAN VALUES: AN ALTERNATIVE APPROACH , 1951 .

[19]  G. S. James.  THE COMPARISON OF SEVERAL GROUPS OF OBSERVATIONS WHEN THE RATIOS OF THE POPULATION VARIANCES ARE UNKNOWN , 1951 .

[20]  Chul H. Ahn,et al.  Modified ANOVA for Unequal Variances , 2003 .

[21]  A. Scott,et al.  Interval Estimates for Linear Combinations of Means , 1971 .