This web page is about an R package (written by Yun Ju Sung and Charles J. Geyer) for doing Logit-Normal generalized linear mixed models (GLMM) using ordinary, independent and identically distributed Monte Carlo.
The package is now hosted at GitHub
https://github.com/cjgeyer/bernor
As the README at GitHub says, the package is easily installed using
R package remotes (https://cran.r-project.org/package=remotes)
library(remotes)
install_github("cjgeyer/bernor", subdir = "package/bernor")
For more info see the package vignette
or the R help files for R functions bnlogl
and bnbigw.
vignette("examples", "bernor")
library("bernor")
help("bnlogl")
help("bnbigw")
A paper about the theory used by this package and using the package for examples is
Sung, Y. J. and Geyer, C. J. (2007).
Monte Carlo likelihood inference for missing data models.
Annals of Statistics, 35, 990–1011.
doi:10.1214/009053606000001389
https://projecteuclid.org/euclid.aos/1185303995
Of no real interest except that the proofs are longer and more detailed is the first draft PDF.
Detailed verification of conditions of the theorems in the paper for the models done by the package.
A redo for the revision of the Booth and Hobert example in the paper.
A new example for the revision from Coull and Agresti.
A redo for the revision of the salamander example, also from Booth and Hobert, data originally from McCullagh and Nelder.