Dear Rizwana and Simon, see also this reply from David Ackerley to a question of mine on r-squared in pGLS
http://www.mail-archive.com/r-sig-phylo%40r-project.org/msg03953.html regards, James > > > Date: Fri, 29 Jan 2016 10:06:05 +1000 > From: Simon Blomberg <s.blombe...@uq.edu.au> > To: r-sig-phylo@r-project.org > Subject: Re: [R-sig-phylo] interpreting pGLS > Message-ID: <56aaaced.3020...@uq.edu.au> > Content-Type: text/plain; charset=windows-1252; format=flowed > > Hi Rizwana, > > There is no reason why we would expect the phylogenetic signal in the > raw variables to be the same as the phylogenetic signal from the > regression analysis (K or lambda), as in the regression, you are really > looking at the phylogenetic signal of the _residuals_, which may be > quite different. Also, the r-squareds from pgls in caper are not > calculated in the same way as r-squareds from an ordinary least squares > analysis. In fact, there is no "correct" way to calculate r-squared for > any model other than the OLS model, as OLS r-squared is based on the > residual variance. For other types of models, including GLS, "residual > variance" is not a well-defined concept. Personally, I don't use > r-squared for any model other than OLS models. There are far better ways > to conduct model criticism. > > Cheers, > > Simon. > > On 28/01/16 15:57, Rizwana Rumman wrote: > > Dear R-sid-phylo list, > > > > I am having some problems to interpret the results of a co-evolution > analysis and I hope you can help me to get my head around things. > > > > I have two variables that each show significant non-random phylogenetic > signal (K and lambda) and appear to be significantly correlated in a > non-phylogenetic regression; but when I am performing a pGLS in caper, the > lambda is estimated to be zero and the two variables shows the exact same > R-square and level of significance as estimated from the non-phylogenetic > regression analysis. I also have two other sets of variables that show > lambdas >0 (very similar values from caper and ape) in the pGLS analysis, > but for these variables, the R-square values are smaller from pGLS than > those estimated from non-phylogenetic regression. I guess my question is if > I can safely interpret that, in the first case, phylogenetic history does > not significantly affect my trait correlations (indicated by lambda=0 and > very similar simple linear regression and pGLS results) and in the latter > case (i.e. smaller R-squares after correcting for phylogeny), > non-phylogenetic regressions slightly overes! > ti! > > mates the correlations of the traits. > > > > Many thanks for the help! > > > > Cheers, > > Rizwana > > > > > > [[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/ > > -- > Simon Blomberg, BSc (Hons), PhD, MAppStat, AStat. > Senior Lecturer and Consultant Statistician > School of Biological Sciences > The University of Queensland > St. Lucia Queensland 4072 > Australia > T: +61 7 3365 2506 > email: S.Blomberg1_at_uq.edu.au > http://www.evolutionarystatistics.org > > Policies: > 1. I will NOT analyse your data for you. > 2. Your deadline is your problem. > > Basically, I'm not interested in doing research > and I never have been. I'm interested in > understanding, which is quite a different thing. > - David Blackwell > > > > ------------------------------ > > Subject: Digest Footer > > _______________________________________________ > R-sig-phylo mailing list > R-sig-phylo@r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-phylo > > ------------------------------ > > End of R-sig-phylo Digest, Vol 96, Issue 17 > ******************************************* > [[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/
