[R-sig-ME] Prediction variance for GLMM

John Maindonald john@m@indon@ld @ending from @nu@edu@@u
Sat Jan 5 21:32:14 CET 2019

You might consider using glmmTMB::glmmTMB(), which has an
`se.fit=TRUE` argument.  Note the ability, with allow.new.levels=TRUE,
to predict for a new random factor level.

glmmTMB() has the advantange of allowing a wider range of error
familiies.  For betabinomial, negative binomial, and other such
families, it allows the modeling of the ‘dispersion’ parameter
(NB that this is not ‘dispersion’ as defined for glm quasi models,
albeit one can be expressed as a function of the other.)

There is the standard warning that the SEs are contingent on the
model assumptions, and on distributional theory approximations.
(E.g., what DF assumptions will give a good approximation to the
coef/SE distributions.)  The lme4 parametric bootstrap function
bootMer() should now work to give SEs that pretty much get
around issues with distributional theory approximations, but not
around model assumptions more generally.

John Maindonald             email: john.maindonald using anu.edu.au<mailto:john.maindonald using anu.edu.au>

On 6/01/2019, at 07:48, Levine, Michael <mlevins using purdue.edu<mailto:mlevins using purdue.edu>> wrote:

Dear all,

I would like to ask the following question. Is it possible to obtain prediction variances for GLMMs in the package lme4 , based e.g. on the function glmer? I know that it is possible to do it with "pure" GLM's but I don't see any options for GLMM's.  I realize there is a problem there because such a variance can be defined in several different ways...

Let me know and thanks a lot in advance!


Michael Levine
Associate Professor, Statistics

Department of Statistics
Purdue University
250 North University Street
West Lafayette, IN 47907 USA

email: mlevins using purdue.edu<mailto:mlevins using purdue.edu>
Phone: +1-765-496-7571
Fax:   +1-765-494-0558
URL:   www.stat.purdue.edu/~mlevins

[[alternative HTML version deleted]]

R-sig-mixed-models using r-project.org mailing list

	[[alternative HTML version deleted]]

More information about the R-sig-mixed-models mailing list