Hello Oliver,
I completely agree with Julien. I’d add however that when running a PGLS with
different branch length transformations you are in fact comparing among
different models of trait evolution, which make different assumptions about how
traits change along the phylogeny. Thus, I would
Hi Oliver,
Your model comparison tells you that the "lambda" is better than "kappa", and
hence that it should provides a better phylogenetic structure for the
corresponding regression test.
Making a model comparison to choose the structure is probably best practice,
but it also depends on the