Meeghan,

As far as I can understand your problem you may not be able to compute 
an overall percent of morphology ( which is not one variable, but 
several) explained by your 5 environmental variables.  What you 
definitely can compute would be a single percentage (r^2) of each 
morphological variable explained by the five environmental variables.

By doing PCA on the morphological variables you might find one major 
PCA that has a sizable percentage of its variability explained by the 5 
environmental variables.  Another might also have much of its 
variability explained.  Adding them together might equal greater than 
100%.   By looking at total variability via summing the PC or looking 
at them individually you can perhaps understand why the fractions 
explained of each can not be added together in any way to give an 
overall % explained.  It would be like adding apples and oranges.

Joe Kunkel

On Sep 8, 2004, at 9:31 AM, [EMAIL PROTECTED] wrote:

> Hello,
> I am a PhD student at the University of Adelaide. I am working on
> morphological data of a common species of marine kelp that I collected
> across Southern Australia (i.e. height, weight, surface area etc). I
> also have 5 environmental variables (latitude, longitude, exposure,
> depth and plant density) and I am wanting to do a canonical
correlation
> analysis (CCA) to examine the overall ability of the suite of recorded
> environmental variables to explain the variation in morphological
> characters.
>
> I have attempted a CCA in numerous programs (Statistica, SYSTAT and
> PC-ORD) but can not get all the information that I require from these
> programs (see query below). Whilst I would like to try the analysis on
> SPSS, this needs to be done on the MANOVA subprogram of SPSS. Firstly,
> is there anyone familiar with this program  is it part of SPSS or do I
> need to access this as a separate program?
>
> Using the 3 programs mentioned above, I have been able to calculate
the
> correlation coefficients for the morphological and environmental
> variables and yet can not calculate the variance explained (%) by each
> canonical axis (i.e. the proportion of the pattern common to the two
> data sets represented by each canonical axis).
>     Secondly, whilst I can obtain the correlations between canonical
> axes and morphological and environmental variables (i.e. canonical
> factor loadings) as well as the variance explained by each axis (for
> both the Morphological and Environmental data sets), I can not 
> calculate
> the cumulative variance of morphological characters explained by
> canonical variates of environmental variables (%) (i.e. what the % of
> the variation in the morphometric data set do the 5 environmental
> variables (collectively) explain). Both of these outputs are the main
> results I am after and I am unable to obtain them. Any advice on how
to
> calculate these results or programs that include these results in the
> CCA output would be greatly appreciated!
>
> Kind Regards,
> Meegan
>
> _____________________________________
>
> Meegan Fowler-Walker
> PhD student
> Southern Seas Ecology Laboratories
> School of Earth and Environmental Sciences
> University of Adelaide
> South Australia 5005
> Australia
>
> Phone: +61 8 8303 6224
> Fax: +61 8 8303 4364
> Email: [EMAIL PROTECTED]
> ==
> Replies will be sent to list.
> For more information see 
> http://life.bio.sunysb.edu/morph/morphmet.html.
>
>
----------------------------------------------------
Joseph G. Kunkel, Professor
Biology Department
University of Massachusetts Amherst
Amherst, MA 01003
http://www.bio.umass.edu/biology/kunkel/
==
Replies will be sent to list.
For more information see http://life.bio.sunysb.edu/morph/morphmet.html.

Reply via email to