Hello everyone,

I work with fluctuating asymmetry of wings and I am currently using 
individual values generated by a Procrustes ANOVA to perform an allometry 
regression and a principal component analysis of asymmetric shape 
variation. I have run the same analyses with datasets of other traits with 
matching symmetry and everything went smoothly. However, in my wing 
dataset, MorphoJ does not allow me to select a fundamental classifier for 
my analyses, which corresponds to the environmental condition (impacted 
site, reference site).

Whenever I try to perform a PCA considering the differences of 
environmental conditions of each individual, the analysis appears on my 
project tree, but with no results. I can perform it without pooling within 
the environmental condition classifier and see the results. Unfortunately, 
a PCA without identifying which condition each individual belongs to would 
not be useful in any way to respond the questions of my research.

I tried to insert the environmental condition effect in the Procrustes 
ANOVA and this message appeared on the reports tab: *"Procrustes ANOVA: The 
effect indicated for Error 2 has no degrees of freedom. The effect was 
dropped from the analysis." . *I found a discussion 
<https://www.mail-archive.com/[email protected]/msg02885.html> on 
Morphmet about this, but I'm not quite sure as to what to do with this 
information. If the issue is related to the error 2 (landmark error), why 
do I only get this warning if I try to include an additional effect on the 
analysis? What can I do to obtain landmark errors with degrees of freedom 
over zero and perform this PCA successfully?

This only happened for analyses of this trait, which has an appropriate 
amount of landmarks according to the sampling size. I have tried to run the 
same analyses with a new set of landmarks and nothing has changed. Is 
anyone familiar with this issue?

Let me know if any more information is needed!

Best regards,
Helena.

-- 
You received this message because you are subscribed to the Google Groups 
"Morphmet" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/morphmet2/8685852c-47be-419c-baf0-6f385d6fe0ccn%40googlegroups.com.

Reply via email to