[R] manova: R vs SAS...need some clarification

Pedro Mardones mardones.p at gmail.com
Tue Aug 12 22:56:32 CEST 2008


Dear all;
working with a 'fat' data set (700 variables / 50 samples) and trying
to run a manova test on it (I'm aware that it's not the best option
for this kind of data set) I got the error in the summary.manova
function about the rank of the residuals (rank < # variables). Ok. The
thing that I don't understand is why I don't get the same type of
error in SAS. There seems to be no problem with rank deficiency and
the fit-statistics in SAS (no negative DF or something like that...).
I'm sure it must be some differences in the way the manova test is
calculated but I don't know what they are, so I'll appreciate any
comments...
Thanks
PM



More information about the R-help mailing list