Skip to content
Prev 15716 / 20628 Next

nAGQ = 0

Oh and on the model comparison thing the only way i know how to directly
compare random effects distribution estimates is with gateaux derivatives a
la what Sophia Rabe-Hesketh did in her GLLAMM software for nonparametric
random effects estimation.

Usually i just kind of eyeball it and try to be overly conservative with
quadrature points or use MCMC.

I guess you could rig up some regular deviance test to do the same thing?
On Sep 3, 2017 6:51 PM, "Poe, John" <jdpo223 at g.uky.edu> wrote: