[R] ML, REML and several random effects

[Gregor]

Arnaud Mosnier <arnaud_mosnier <at> UQAR.QC.CA> writes:
> Hello everyone,
> 
> Hope that my question could have interest for more than one people here.
> I know that in order to compare mixed models with different fixed effect 
> I need to use ML method.
> But ... what about comparing models with the same fixed effects and 
> different random effects ?

Likelihood ratio test on REML likelihood

Gregor