Abstract
Classical multivariate analysis is based on the assumption that the data come from a multivariate normal distribution. The tests of multinormality have therefore received very much attention. Several tests for assessing multinormality, among them Mardia’s popular multivariate skewness and kurtosis statistics, are based on standardized third and fourth moments. In Mardia’s construction of the affine invariant test statistics, the data vectors are first standardized using the sample mean vector and the sample covariance matrix. In this paper we investigate whether, in the test construction, it is advantageous to replace the regular sample mean vector and sample covariance matrix by their affine equivariant robust competitors. Limiting distributions of the standardized third and fourth moments and the resulting test statistics are derived under the null hypothesis and are shown to be strongly dependent on the choice of the location vector and scatter matrix estimate. Finally, the effects of the modification on the limiting and finite-sample efficiencies are illustrated by simple examples in the case of testing for the bivariate normality. In the cases studied, the modifications seem to increase the power of the tests.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
A. Bera and S. John, Tests for Multivariate Normality with Pearson Alternatives. Commun. Statist. Theor. Meth. 12 (1983), 103–117.
C. Croux, G. Dhaene, and D. Hoorelbeke, Testing the Information Matrix Equality with Robust Estimators. Submitted, 2003, available at http://www.econ.kuleuven.ac.be/public/NDBAE06/public.htm.
C. Croux and G. Haesbroeck, Principal Component Analysis Based on Robust Estimators of the Covariance or Correlation Matrix: Influence Functions and Efficiencies. Biometrika 87 (2000), 603–618.
F. R. Hampel, E. M. Ronchetti, P. J. Rousseeuw, and W. A. Stahel, Robust Statistics: The Approach Based on Influence Functions. Wiley, New York, 1986.
G. Koshevoy, J. Möttönen, and H. Oja, A Scatter Matrix Estimate Based on the Zonotope. Ann. Statist. 31 (2003), 1439–1459.
J. A. Koziol, Probability Plots for Assessing Multivariate Normality. Statistician 42 (1993), 161–173.
K. V. Mardia, Measures of Multivariate Skewness and Kurtosis with Applications. Biometrika 57 (1970), 519–530.
K. V. Mardia, Applications of Some Measures of Multivariate Skewness and Kurtosis in Testing Normality and Robustness Studies. Sankhya, Ser B 36 (1974), 115–128.
K. V. Mardia, Tests of Univariate and Multivariate Normality. In P. R. Krishnaiah, editor, Handbook of Statistics, vol 1, pages 279–320. North Holland, 1980.
J. Möttönen and H. Oja, Multivariate Spatial Sign and Rank Methods. J. Nonparam. Statist. 5 (1995), 201–213.
H. Oja, Descriptive Statistics for Multivariate Distributions. Statist. Probab. Lett. 1 (1983), 327–332.
E. Ollila, T. P. Hettmansperger, and H. Oja, Affine Equivariant Multivariate Sign Methods. Submitted, 2003a.
E. Ollila, H. Oja, and C. Croux, The Affine Equivariant Sign Covariance Matrix: Asymptotic Behavior and Efficiencies. J. Multivariate Anal. 87 (2003b), 328–355.
S. Visuri, V. Koivunen, and H. Oja, Sign and Rank Covariance Matrices. J. Statist. Plann. Inference 91 (2000), 557–575.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer Basel AG
About this paper
Cite this paper
Kankainen, A., Taskinen, S., Oja, H. (2004). On Mardia’s Tests of Multinormality. In: Hubert, M., Pison, G., Struyf, A., Van Aelst, S. (eds) Theory and Applications of Recent Robust Methods. Statistics for Industry and Technology. Birkhäuser, Basel. https://doi.org/10.1007/978-3-0348-7958-3_14
Download citation
DOI: https://doi.org/10.1007/978-3-0348-7958-3_14
Publisher Name: Birkhäuser, Basel
Print ISBN: 978-3-0348-9636-8
Online ISBN: 978-3-0348-7958-3
eBook Packages: Springer Book Archive