Pseudo-AIC and GLIMMIX

[deleted_user]

Hi

I realise that this might be the wrong SAS forum, but I didn't find a header that I felt suited this topic.

I'm using proc glimmix and is having problems with model selection, and the reported Fit statistics. My initial idea was to use AIC to choose between competing models, but after I read the (short) info on pseudo-AIC I feel more hesitiant.

The SAS printout states:
"Fit statistics based on pseudo-likelihoods are not useful for comparing models that differ in their pseudo-data."

And in the glimmix help file you can read:
"Note that the (residual) log pseudo-likelihood in a GLMM is the (residual) log likelihood
of a linearized model. You should not compare these values across different
statistical models, even if the models are nested with respect to fixed and/or G-side
random effects. It is possible that between two nested models the larger model has a
smaller pseudo-likelihood."

My interpretation is that you cannot use the pseudo-AIC to compare e.g. two nested models:
model resp = k range;
model resp = k|range;

I've tried to find more information on the pseudo-AIC, but without much results. Is my interpretation correct? If so, can the pseudo-AIC still be used to evaluate and choose between different correlation structures in the residuals (e.g. sp(exp) vs AR(1)) ?

I would be very thankful for all comments.

[Damien_Mather]

Do you absolutely have to model R-side effects? Maybe you can control for them another way? Maybe a fixed effect? If so then you can specify method=laplace or method=quad(qpoints=n). If you have large data and lots of random effects then stick to laplace or quad(qpoints=1) equivalent to start wtih.

[Cynthia_sas]

Hi:

  There is a SAS Statistical Procedures forum

http://communities.sas.com/community/sas_statistical_procedures

  And, GLIMMIX would seem, to me, to fall under that category. There are WAY more statistical folks hanging out over in that forum than in the ODS forum.

  Just my .02,

cynthia