You might want to rethink about getting model averaged coefficients. That
is a bunch of nonsense if you have any multicollinearity among the
predictors. Model averaged predictions might be useful.
Brian
Brian S. Cade, PhD
U. S. Geological Survey
Fort Collins Science Center
2150 Centre Ave., Bl
Hello,
I am using a negative binomial distribution in glmmADMB to fit a mixed model
and then using the MuMIn package to get model averaged coefficients. As far as
I can tell, this approach gives no estimates for the variance of the random
effects. I have been taking these from the top model a
2 matches
Mail list logo
