www.cmm.bristol.ac.uk/forum

Hi,

I have successfully run a number of models using a continuous outcome variable, with varying degrees of complexity - up to 5 levels and over 30 fixed effects (in MLwiN 2.26). However, now I am using the same dataset to run a binomial, logistic model, I am able to specify the equation, but when I try to start the IGLS estimation I get this error message:

"error while obeying batch file D:\Program Files (x86)\MLwiN
v2.26\i386\..\discrete\pre at line number 11:
name c1180 'H' C1181 'F~(H)' C1182'H*' C1183 'P' c1184 'MASK' c1189 'Y-VAR'

Duplicate column name."

Strangely, if I skip the IGLS estimation, and go straight to MCMC (bi-passing the warning re: '0' values) it works, but I'm worried that by running MCMC without previously running IGLS I will be getting inaccurate estimates.

Any ideas on how to resolve this error?

Thanks,

Caroline.

I have now copied the entire dataset into a new MLwiN worksheet and it seems to be working.

I'm glad to hear that you now have this working. An alternative fix would have been to check for columns with the names mentioned in the error message and then remove/rename these.