No, they are the same. The difference between GLS and OLS is only in the specified (assumed) variance-covariance matrix of the residuals. Cheers Ted P.S. "PGLS" can be viewed as costing 1 df if you estimate a branch length transformation parameter.
On Fri, Jun 11, 2021, 9:54 PM Russell Engelman <neovenatori...@gmail.com> wrote: > Dear R-Sig-Phylo Group, > > I had a question regarding the calculation of regression statistics using > PGLS. When I obtain the degrees of freedom from a PGLS object using ape in > R, I notice that it returns the same number of degrees of freedom as under > OLS, and the AIC of a bivariate function is calculated under the assumption > that k=2 (i.e., that the same number of degrees of freedom are used as in > the OLS). Does anyone know why this is? I would have thought that a PGLS > would have eaten up a greater number of degrees of freedom than OLS, if for > no other reason than it constrains the motion of the system by adding > phylogenetic covariation in as an additional variable, and thus would be > expected to eat up at least one additional degree of freedom in the same > way that adding additional categorical or quantitative variables would? > > Sincerely, > Russell > > [[alternative HTML version deleted]] > > _______________________________________________ > R-sig-phylo mailing list - R-sig-phylo@r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-phylo > Searchable archive at > http://www.mail-archive.com/r-sig-phylo@r-project.org/ > [[alternative HTML version deleted]] _______________________________________________ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
