Dear list members,
I have an acute form of analysis paralysis. Could please somebody give
me a push into the right direction?
I�m trying to unravel which environmental / habitat variables are the driver
of species composition. It�s a huge data set, with 70 plus environmental
variables and a large arthropod data set. So, I will do the same type of
analysis for different arthropods subgroups.
First, I did a PCA to reduce the number of env. variables (still 30
variables) followed by a RDA (rda) on Hellinger transformed data and
stepwise model selection (ordiR2step and by hand). I standardized the
environmental variables beforehand (scale), but I�m not sure if this is not
already included in the rda function?
However, the model result changes a lot if I change the input order of the
variables. I know this is normal but it feels like a rather subjective
approach, especially as some of the variables seem to me equally important
and for others I have no idea how they might affect species composition�
I read about the bioenv function � is this a better way to go?
I�m very thankful for any comments, ideas etc.,
Cheers,
Conny
[[alternative HTML version deleted]]
_______________________________________________
R-sig-ecology mailing list
R-sig-ecology@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
