[R-sig-ME] Suggestions on how to correct a

[Bailey, Paul]

Hi Martin,

I don't know how to reply to the digest and get the thread right, so I apologize if I get it wrong.

I think the 2015 paper you link to describes linear models but you say "lme4" below. I think you mean only lme4::lmer.

The vignette labeled "Computational Methods" in the CRAN package (v 1.1-31) in section 4.2 "The PIRLS algorithm for [u-tilde] and [beta-tilde]" says that this is what is done for glmer as well as lmer. But I think that can't be, because I think that would then mean that when I requested the Laplace approximation (nAGQ=1) I would be getting the PLS results (nAGQ=0 results) for beta. But please correct me if I'm wrong.


Best,
Paul Bailey