[R] nested unbalanced ANOVA

Lorenz Gygax lgygax at access.unizh.ch
Fri Apr 16 15:25:34 CEST 1999

Dear all,

is there a standard way to calculate a nested, unbalanced ANOVA with R?
I am glad for pointers, tips, ideas.

The problem that I am trying to evaluate is a comparison of (average) 
group sizes. The explanatory variables that I would like to include are
species which are nested within families and (partly) crossed with region,
i.e. some species occurr in several regions while others do not. For some
species I have several estimates of average group size in a given region
and, of course, not all families have the same number of species but each 
species only occurrs in one family.

I actually went back to Splus (5.0 R2 for Sun SPARC) because Splus has the
nested %in%, rsp. / which can be included in a formula, but the two models

res <- aov (gs ~ fam/spe + reg)
res <- aov (gs ~ fam + reg + spe)

calculate exactly the same ANOVA table. Am I completely wrong or they?

Is it ok instead, if I calculate aov (fam + spe + reg) but test the family
in that I divid its mean square by the mean square of species to get an
F-value or is this too crude? 

Thanks for your help. Regards, Lorenz
--			~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Lorenz Gygax		 LGygax at amath.unizh.ch;       room: 36-L-40
			 Department of Applied Mathematics
			 University of Zuerich-Irchel
			 Winterthurerstr. 190; CH-8057 Zurich
			 voice: 41-1-635-58-52  fax: 41-1-635-57-05

r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list