[R] multivariate normality test

Remigijus Lapinskas remigijus.lapinskas at mif.vu.lt
Thu Jun 29 16:37:44 CEST 2006


Hello,

Could someone help me to explain the VERY big difference in applying two
tests on multivariate normality:

library(mvnormtest)
data(EuStockMarkets)
mshapiro.test(t(EuStockMarkets[15:29,1:4]))


         Shapiro-Wilk normality test
data:  Z
W = 0.8161, p-value = 0.005955

and

library(energy)
mvnorm.etest( EuStockMarkets[15:29,1:4] )

         Energy test of multivariate normality: estimated parameters

data:  x, sample size 15, dimension 4, replicates 999
E-statistic = 1.0041, p-value = 0.2482



Many thanks,
Rem



More information about the R-help mailing list