Error 1909 using MCMC on Multivariate data

Welcome to the forum for MLwiN users. Feel free to post your question about MLwiN software here. The Centre for Multilevel Modelling take no responsibility for the accuracy of these posts, we are unable to monitor them closely. Do go ahead and post your question and thank you in advance if you find the time to post any answers!

Remember to check out our extensive software FAQs which may answer your question: http://www.bristol.ac.uk/cmm/software/s ... port-faqs/
Post Reply
thomasm
Posts: 1
Joined: Tue Mar 29, 2011 12:01 pm

Error 1909 using MCMC on Multivariate data

Post by thomasm »

Hi
as a newcomer to the forum I'd welcome any suggestions regarding the following error message:
"MCMC Error 1909: Currently missing X variables must be listwise deleted in multivariate models for the MCMC methods (MCMC)"

Using educational data, I am modelling 5 plausible values in a multivariate structure, which makes it a 3-level model with response-student-school structure.
My model was set up with only one additional variable that had some missing values. Having used IGLS estimates to provide appropriate starting values, I had hoped that on changing the estimation method to MCMC I would have been able to fit the model again to get imputed values for the missing Xs and then to repeat the process to calculate and store an imputed dataset. [As suggested in Chapter 16 of Browne's MCMC estimation in MLwiN Version 2.10]

Unfortunately the process stops before I get there, with Error 1909 as cited above.
If the X variables are listwise deleted, then how can the model include those contributions, as I would be effectively reducing the data to complete sets only?

On a more general note, where are the descriptors stored for the various error codes in MLwiN?
billb
Posts: 163
Joined: Fri May 21, 2010 1:21 pm

Re: Error 1909 using MCMC on Multivariate data

Post by billb »

Hi Thomasm,

If you wish to create imputed datasets to fill in what will ultimately be X variables in your model to fit you need to have them as response variables in the imputation model you first fit.

Say for example with the hungary dataset your final model is going to be a regression of biol_core with predictors biol_r3 and biol_r4 then you could first fit the models as described in the manual to create fully imputed datasets, then split these columns up to get columns of the correct length and then fit the models.

I hope this is of some help,

Bill.
Post Reply