Summary
Mardia’s coefficients of multivariate skewness and kurtosis can be used to assess the multivariate normality assumption that must be satisfied in many multivariate statistical procedures. However, the asymptotic tests of multivariate skewness and kurtosis do not perform well in small samples. A Monte Carlo method for accurately estimating the p-value is proposed and illustrated.
Similar content being viewed by others
References
Berry, D.A. & Lindgren, B.W. (1996), Statistics: Theory and Methods, 2nd ed. Duxbury, Belmont.
D’Agostino, R.B. & Stephens, M.A. (1986), Goodness-of-fit Techniques. Marcel Dekker, New York.
Mardia, K.V.(1970), Measures of multivariate skewness and kurtosis with applications, Biometrika, 57, 519–530.
Mardia, K.V. (1974), Applications of some measures of multivariate skewness and kurtosis for testing normality and robustness studies,” Sankhya B, 36, 115–128.
Rencher, A.C. (1995), Methods of Multivariate Analysis. Wiley, New York.
Seber, G.A.F. (1984), Multivariate Observations. Wiley, New York.
Author information
Authors and Affiliations
Appendix: Gauss program
Appendix: Gauss program
Rights and permissions
About this article
Cite this article
Bonett, D.G., Woodward, J.A. & Randall, R.L. Estimating p-values for Mardia’s coefficients of multivariate skewness and kurtosis. Computational Statistics 17, 117–122 (2002). https://doi.org/10.1007/s001800200094
Published:
Issue Date:
DOI: https://doi.org/10.1007/s001800200094