From: John Gonzales on
Hi,

There are numerous examples for using model selection (AIC, etc.) to select
the best covariance structure for "proc mixed" models.

However, I am interested in ranking models with different fixed effects (not
nested). It is my understanding that the REML procedure cannot be used for
model selection unless the models are nested. If I use maximum likelihood
(ML), the results for the fixed effects can differ considerably from the
same model estimated via REML (i.e. same covariance structure and model
statement).

Is there any procedure for ranking models with non-nested fixed effects
using ML? Most examples use the REML procedure, and I am not sure if ML is
considered appropriate. I have searched around, but almost all of the model
selection examples are for determining the best covariance structure. I am
interested in comparing the competing models themselves.

cheers.