[R-sig-ME] interaction model and variance in lme

David Springate David.Springate at postgrad.manchester.ac.uk
Wed Feb 11 23:19:42 CET 2009


I am new to R and I have been trying to build a model that I can  
extract ML variance components for the interaction from, but am  
struggling to make sense of the formula.

I am looking at the interaction of family (random) and two  
environmental treatments (fixed) on a trait.  So far I have tried:


But this seems to give me just the treatment nested in the family  
rather than family*treatment values.

I also tried:

Model = lme(trait~family*treatment,data=dataset,random=~1|family)

which gives me an interaction MS and F value in an anova, but seems to  
treat each interaction between each treatment and family as a separate  
fixed factor.  Also VarCorr() doesn't seem to give a variance for the  
interaction term.

What am I doing wrong?

I am sure this should be simple, but the docs seem pretty unclear to  
me on modeling interactions.

Help please!

David Springate

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