[R] gls anova wald test calculations

Alan Swanson swanson at yellowstoneresearch.org
Tue Mar 2 20:18:43 CET 2004

I have a question about the Wald test F-statistics that are calculated when 
the anova() command is used on a singular gls or lme object.  As I recall 
from my linear models class, the Wald test examines H0: C'B = d0 vs Ha: C'B 
!= d0.  Does anybody know how this C matrix is constructed in R?  Is there 
a way to see the C matrix that R is using?  In my situation, I'm looking at 
marginal ANOVA tables and would like to know how the other factors, and in 
particular the interaction terms, are being treated for a given row of the 
table.  I'm assuming that some sort of averaging takes place but can't find 
anything about it in the documentation I have.  Thanks, Alan

More information about the R-help mailing list