logistic multilevel error

[jelliesierksma]

Hi,

I keep on getting an error while performing a logistic multilevel model with mlwin:

error while obeying batch file Q:\mlw221|discrete/post at line number 30 qdivi 'F ~ (H)'

error in aritmethic expression

Anybody know what this is all about?

Thanks in advance!

Jellie

[ChrisCharlton]

Without seeing your model/data I can't give a definite reason for this error message. What linearisation method are you using? You will sometimes see messages like this when it is having trouble fitting using second order PQL. If this is the case you can sometimes get the model to converge by first fitting with first order MQL and then changing to second order PQL and using the "more" button. Another alternative is to fit the model using MCMC.

[jelliesierksma]

Thanks for your reply.

It concerns the following:
binary outcome, two levels (responses in children), bionomial districbution with logit link.

When I run just this model, I get the error message. Adding anything to the model does not work. I was planning on adding two dummy variables to it.

Also, If i run his particular model on another computer the error does not occur.

Any help would be greatly appreciated!

Kind regards,
Jellie

[ChrisCharlton]

Have you checked whether the version of MLwiN is the same on both machines? If it isn't then what are the version numbers of the one that works and the one that doesn't?

[jelliesierksma]

Actually, it is the same version (only the computer is different), namely 2.21
It is very strange...

Perhaps, I new computer is the solution.

[jelliesierksma]

i also used the default linearisation, MQL 1st order.. i don't think that is the problem either.

[ChrisCharlton]

That's a bit of a mystery then. I would suggest that you update to the most recent version of MLwiN (http://www.bris.ac.uk/cmm/software/mlwi ... rades.html) in case this was caused by something that has been fixed since version 2.21. If the error message is highlighting the correct location then the GDIV command is attempting to divide all the columns in G1 (see Data Manipulation -> Group for the contents of this) by a column called 'F~(H)'. You could check that G1 contains a list of columns and that the 'F~(H)' column has been created, however if this works on another machine then this should be the case.