Dear all,
I have a question conserning the pgls regression in package caper. The function allows to estimate or fix three branch length transformations. I wanna figure out which transformation gives me the best model fit by comparing for example a model with lambda estimated (lambda=ML) to a
Hi Sereina.
Why lambda=0.5? Normally investigators tend to compare a model where
lambda is estimated to one in which it is fixed at 1 which corresponds
to Brownian evolution; or 0 which corresponds to no phylogenetic
correlation in the residual error of the model.
We can compare two fitted
Thank you guys for you helpful advice!
The probelm in my model is that it estimates lambda to be 1, however, my gut feeling would say this is rather unlikely. The model where lambda equals 1 looses significance completely, however, with lambda fixed at 0.85, it is highly significant. So what
