[R] likelihood ratio test between glmer and glm

[Ben Bolker]

Davnah Urbach <Davnah.Urbach <at> dartmouth.edu> writes:


> Thanks for this answer but does that mean that working 
> with the deviances is better? Or how else could I
> evaluate the importance of my random terms?

  You should probably (a) search the archives of the 
r-sig-mixed-models mailing list
and (b) ask this question again there.