On Tue, 12 Aug 2008, Gareth Campbell wrote:
Hmm, I have my elements (as in chemical elements) as columns and my samples
as rows. This is the normal way round for all my other analyses. I think I
figured it out... I have 6 groups, all of about 10-12 samples and each with
31 elements. So for each group there are many more elements than samples
and it makes the analysis impossible due to the problems of the singular
covariance matrix, if I've got that right.
Is that the problem?
If this is a MANOVA with 31 responses (elements) then the problem is that
those responses are linearly dependent after subtracting group means (and
quite possibly before). With 86 samples this is not normal (MANOVA is
using a common covariance matrix for each group, not one for each group).
I do wonder what your aim is: this seems to be using MANOVA to do LDA.
2008/8/12 S Ellison <[EMAIL PROTECTED]>
Rank deficiency is usually an indication of under-determination, if I've
got it right.
But something is clearly odd, because with 31-column matrix of 86 rows,
you should have 86 residuals for each of 31 models, not 31 residuals.
Could your matrix be the wrong way round?
"Gareth Campbell" <[EMAIL PROTECTED]> 11/08/2008 23:38 >>>
Thanks for that, the error I get when I use manova is this:
Error in summary.manova(fit) : residuals have rank 30 < 31
What does this mean??
When I call the formula I get 31 residuals and 31 residual standard
errors.
Not sure why I get this error?
Thanks
Gareth
2008/8/11 S Ellison <[EMAIL PROTECTED]>
test="Wilks" is available in summary.manova, but not in summary.aov;
could that be the problem? I find in ?summary.manova the example
makes a
clear distinction between these three summaries...
fit <- manova(Y ~ rate * additive)
summary.aov(fit) # univariate ANOVA tables
summary(fit, test="Wilks") # ANOVA table of Wilks' lambda
summary(fit) # same F statistics as single-df
terms
Another common variable in anova using R is the way contrasts are
set;
you might check that with your colleague?.
Steve E
"Gareth Campbell" <[EMAIL PROTECTED]> 08/10/08 11:43 PM >>>
Hi,
I'm doing anova on a matrix of multivariate data where I want to
assess
the
effect of each column (element).
My matrix is 86 rows x 31 columns. I've created a grouping factor
of
length
86 containing group assignments of 6 types.
Then I run:
x<- aov(matrix~grouping.factor)
summary(aov.fit.raw, test="Wilks")
This is working fine enough, but I'm getting different results to
someone
who I'm comparing with - am I doing what I think I am doing here??
What I think I'm doing is - take the first column (element) and then
look at
that element between the 6 groups and report F, Pvalue etc..., then
move
onto the 2nd column and repeat.
Thanks
--
Gareth Campbell
PhD Candidate
The University of Auckland
P +649 815 3670
M +6421 256 3511
E [EMAIL PROTECTED]
[EMAIL PROTECTED]
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
*******************************************************************
This email and any attachments are confidential. Any use, copying or
disclosure other than by the intended recipient is unauthorised. If
you have received this message in error, please notify the sender
immediately via +44(0)20 8943 7000 or notify [EMAIL PROTECTED]
and delete this message and any copies from your computer and
network.
LGC Limited. Registered in England 2991879.
Registered office: Queens Road, Teddington, Middlesex, TW11 0LY, UK
--
Gareth Campbell
PhD Candidate
The University of Auckland
P +649 815 3670
M +6421 256 3511
E [EMAIL PROTECTED]
[EMAIL PROTECTED]
*******************************************************************
This email and any attachments are confidential. Any u...{{dropped:24}}
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
--
Brian D. Ripley, [EMAIL PROTECTED]
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
