[R-sig-ME] glmer p-values?

[John Maindonald]

You are of course talking about cases, notably poisson
or binomial, where there is a theoretical value for the
exponential family component of variances.

p-values returned by glmer do suffer from the same issues, 
except where observation level random effects can be 
assumed zero, and then only for parameters (if any) where 
the only contribution to the standard error is from the 
binomial or Poisson variance.  (In this situation, for 
purposes of getting estimates and SEs for those 
parameters, a glm model would give very nearly the 
same result.)

They do not however suffer quite so badly, because
part of the variance (the binomial or Poisson part)
is unaffected by degrees of freedom considerations.

John Maindonald             email: john.maindonald at anu.edu.au
phone : +61 2 (6125)3473    fax  : +61 2(6125)5549
Centre for Mathematics & Its Applications, Room 1194,
John Dedman Mathematical Sciences Building (Building 27)
Australian National University, Canberra ACT 0200.
http://www.maths.anu.edu.au/~johnm

On 02/12/2010, at 1:42 PM, Lucas Kid wrote:

> Am I correct in assuming that the p-values returned by glmer do not suffer
> from the same issues as those that are not returned by lmer (addressed quite
> famously before by Dr. Bates)?
> 
> Thanks!
> 
> Luke
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models