[R-sig-ME] GLS models question

Douglas Bates bates at stat.wisc.edu
Thu Mar 12 18:51:29 CET 2015

It is better to send questions like this to the
R-SIG-Mixed-Models at R-project.org mailing list, which I am cc'ing on this
reply.  Several of those who read that list are better able to respond
quickly than I am.

On Thu, Mar 12, 2015 at 7:55 AM Sophie Elliott <
s.elliott.1 at research.gla.ac.uk> wrote:

> Dear Dr Pinheiro and Dr Bates,
> I’m a PhD student from the university of Glasgow and have been using your
> nlme package in R for data analysis. Specifically I have been using a GLS
> model distribution for my data since I have heterogeneity in my data. I was
> wondering whether it was valid to add an independent random effect to my
> model in addition to the variance structures i.e. I am interested in
> understanding how substrate type and standardised fauna dominance affects
> the standardised abundance of juvenile cod. The data has been standardised
> according to the area of a transect instead of adding an offset since my
> sample size is small due to logistical problems in collecting the data.
> Through model selection procedure I have fitted a GLS with a variance
> structure for fauna dominance, however I would like to add a random effect
> for Julian date and Cluster locations (to take into account temporal and
> spatial variation). Do you think this is possible, if so how could I apply
> this? I have tried applying a random effect and a variance structure a
> normal GLMM using nlme, however the residuals are far more heterogeneous
> when I use a GLS model structure. I also tried adding JD to the GLS
> variance structure of the model only but I wasn’t sure whether this was a
> statistically sound method to carry out the analysis
> e.g. of the GLS I am not sure about:
> VCb<- varComb(varIdent(form=~ 1|factor(JD)),
>               varPower(form=~ IDominance))##power
> GLS4 <- gls(SNumber ~ Substrate + IDominance,
>             weights=VCb,
>             data=Cod)#
> Thank you so much for you time
> Sophie
> PhD Student
> University of Glasgow
> Institute of Biodiversity, Animal Health and Comparative Medicine
> Graham Kerr Building, University of Glasgow, Glasgow G12 8QQ
> Email: S.Elliott.1 at research.gla.ac.uk

	[[alternative HTML version deleted]]

More information about the R-sig-mixed-models mailing list