st: bias in -gllamm- results

["Michael Burton" <mburton@fnas.uwa.edu.au>]

Dear all,

I have been conducting some monte carlo simulations with -gllamm-
models: in particular a probit regression with covariate measurement
error, along the lines of chapter 6 in the manual. I am using Stata
version 9.2

I am finding a consistent divergence in retrieving the known parameter
values when the size of the measurement error increases.