Variance- Covariance matrix

[hdilk0507]

I use MCMC estimation to fit a logistic regression model for a binary response. When i add a level 2 variable as a random effect to the model with level 1 variables and run IGLS, the variance co-variance matrix appeared to be with lot of zeros. Therefore I cannot implement MCMC estimation. How can I fix this problem? Do I have to ignore the level 2 variable?

[ChrisCharlton]

If you're getting all zeros for the level 2 covariance matrix then it's worth checking the hierarchy viewer to make sure that it matches what you expect. If not then your data may not be sorted by the model hierarchy. Assuming that it does look as expected you can choose alternative starting values for this matrix by editing the values in column c1096. As you edit this column the equation windows should update with the new values so that you can see which row corresponds to each element of the random part matrices.

[hdilk0507]

Thanks a lot. I'll follow your instructions. Is it mentioned in the manual to change values in the column c1096? Or did you give that advice depending on any research paper? Please include a reference for that as I have to mention the reason for changing values in my thesis. Actually I have to clarify each and every step I keep. Thanks again.

CHEERS!!

[ChrisCharlton]

Manually setting the starting values is covered in chapter 5 of the MCMC manual (see http://www.bris.ac.uk/cmm/software/mlwi ... ection.5.5).

[hdilk0507]

Thank You.