[R-sig-ME] glmer Z-test with individual random effects

[Andrew Dolman]

Dear Ben,

Can I just check, is it necessary to have both a large total sample
size and a large number of levels of your random effect(s) for a
likelihood ratio test to be robust? Or does the large number of levels
requirement apply only to the Wald test? I'm referring to the part of
your answer below.

Thanks,

Andy.


"They are equivalent to assuming an infinite/large 'denominator degrees
of freedom'.  If you have a large sample size (both a large number of
total samples relative to the number of parameters, and a large number
of random-effects levels/blocks) then this should be reasonable -- if
not, then yes, the 'usual problems with figuring out the number of
parameters' is relevant.  On the other hand, if you're willing to assume
that the sample size is large, then likelihood ratio rests
(anova(model1,model2)) are probably better than the Wald tests anyway."