A matrix of _all_ the relative warps contains the same information as a matrix of partial warp scores. It just represents a change of variables. Therefore a MANOVA, multivariate regression, etc. will give the _same_ statistical results - not better or worse. The difference will be that they are expressed in terms of different variables (partial warps vs. linear combinations of the partial warps). One advantage of the use of relative warps would be that if sample sizes are small then there may be fewer relative warps than partial warps. This may make it possible to invert a covariance matrix in certain analyses - but if sample sizes are that small then you probably should not be using those methods.
I give some discussion of this point in my 1999 paper in the Journal of Classification (16:197-223). ------------------------------------------------- F. James Rohlf -SUNY Stony Brook, NY 11794-5245 FAX: 1-631-632-7626 www: http://life.bio.sunysb.edu/ee/rohlf > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of angela roggero > Sent: Monday, June 07, 2004 9:15 AM > To: [EMAIL PROTECTED] > Subject: Advice on Relative Warps > > Dear morphometricians, > I hope that you could help me to solve a problem which is > bothering me. > Someone tells me that relative warps can be employed instead > partial warps (and with better results) in statistical > analyses, as discriminant analysys and multivariate. I'm not > confident at all to apply relative warps in such way, and I'm > oriented to reject steadily the suggestion. > Moreover, I would like to know also the opinions of the > members of the list. > Thank you very much in advance for any piece of information > and suggestions you could furnish me with. Angela Roggero > > -- > Dott. Angela Roggero > Dpt. Biologia Animale > Via Accademia Albertina 17 > I-10123 Torino - ITALY > Phone +39 011 6704710 > Fax +39 011 6704732 > http://www.dba.unito.it/pers/roggero.html > > > > > _____________________________________________________________________ > For your security, this mail has been scanned and protected by Inflex > == > Replies will be sent to list. > For more information see > http://life.bio.sunysb.edu/morph/morphmet.html. > == Replies will be sent to list. For more information see http://life.bio.sunysb.edu/morph/morphmet.html.
