www.cmm.bristol.ac.uk/forum

Hi there,

I'm using runmlwin in STATA to run a simple multinomial logistic regressions, but MLwinN stops working (message 'error while obeying batch file') and I get a message in STATA "The model did not run properly in MLwiN. Re-run the model without the nopause option to debug the model in MLwiN". I've tried this, but MLwinN keeps crushing, even after having increased the burning and chain. I would like this simple model to work so I can add other predictors to it.

Thanks in advance for your help!

Below is my syntax:

**Generating a multinomial aawlktot variable (aawlktotgrp)
gen aawlktotgrp = aawlktot
recode aawlktotgrp 0=1 1/59=2 60/149=3 150/max=4
label var aawlktotgrp "Total walking mins"
label define aawlktotgrplab 1 "None" 2 "0-59mins" 3 "60-149mins" 4 "150+mins"
label values aawlktotgrp aawlktotgrplab

quietly runmlwin aawlktotgrp cons, level2(cd2006:cons) level1(id:) discrete(distribution(multinomial) link(mlogit) denominator(cons) basecategory(1)) nopause
runmlwin aawlktotgrp cons, level2(cd2006:cons) level1(id:) discrete(dist(multinomial) link(mlogit) denom(cons) basecategory(1)) mcmc(burnin(1500) chain(140000)) initsprevious nopause nogroup
estimates store Model0

Can you post the output from the IGLS part of your model fit (without the quietly prefix)?