Date:         Wed, 7 Mar 2001 04:25:06 GMT
Reply-To:     PD <sophe@USA.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         PD <sophe@USA.NET>
Organization: Icj808
Subject:      Re: multivariate normality test ?

You probably want to visit SAS online sample library. I remember there was a
sample code that covered multivariate normality test. It specifically
mentioned Mardia.

PD
"yogab" <yogab@hotmail.com> wrote in message
news:3AA01EF8.8F6858D4@hotmail.com...
> Hello all.,
>
> Iam looking for some information about  the mutinormality code
> on 'sas-contributors' site 'multinorm.sas'. has any one used it before ?
>
> I was using this test normality (Using - Mardia 1974 method) of some
> data(1000 samples), it
> seem to give an acceptable level of 'p' value for multinorm-kurtosis but
> consistently
> fails for multinorm-skewness also it passes all the univariate test
> (well which does not
> mean they are multinorm, but does not prove otherwise too). I little
> skeptical about this test.
>
>  Has any one used this test before ? Any pointer of consistency or
> limitations of
>  using this code/test.? Has any one used or know any consistent test for
> multi-norm ?
>
>
> thanks in advance.
>
> regards,
>
> yoga@tamu.edu
>
>
>
>
>