Skip to content
Prev 9918 / 20628 Next

Low intercept estimate in a binomial glmm

Hi,

plogis(-2.3776295) is the mode not the mean.

An approximation for the mean is:

c2<-((16*sqrt(3))/(15*pi))^2

plogis(-2.3776295/sqrt(1+c2*4.6432))

and this should be closer to the observed mean.

Cheers,

Jarrod



Quoting Zack Steel <zacksteel at gmail.com> on Wed, 3 Apr 2013 12:58:46 -0700: