On Wed, Apr 7, 2010 at 9:25 PM, Eric Scott <ersco...@illinois.edu> wrote:
> Thank you for your reply. The WoodEnergy example helped a lot. I > understand now that it is inappropriate to make all pairwise comparisons > with an interaction present and better to make comparisons between levels of > one factor within a constant level of the second factor. As I understand it, > the solution in the WoodEnergy example is to produce separate ANOVAs for > each type of wood and then perform the multiple comparisons between stove > types within each wood type. This makes a lot of sense. For my data, I'm > using glm.nb and if I run the model separately for each level of "site," it > estimates a different theta for each which I immagine is a problem. Is this > ok, or do I need to find a way to use the coefficients from the full model > with the interaction to compare levels of clipping within fixed levels of > site? > > -Eric Scott > > The "right" solution is to fit one model and then work with its coefficients. For this example the R glht function did not, at the time I wrote the example, have the option of averaging over the wood types. It now has "experimental" options for interaction_average covariate_average These usually, but not always, do the right thing. For this example, I prefer the analysis in file HH/demo/MMC.WoodEnergy.s.R in which one aov model is calculated and the adjustments are made for the levels of Wood. That file works in S-Plus, but not in R. As I noted before, I still need to revise the WoodEnergy example to use the experimental option in glht to duplicate the results I get from S-Plus. [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.