How to transform repeated measurements into G-side effect

[coollord]

For methods like quad and laplace to work, is there a way to transform repeated measures data into a G-side effect?

Proc glimmix data = blah plots=residualpanel;

Class a1 a2 a3;

Model ff = a1/ dist = poisson  s;

random a2/ subject = a3 type = cs;;

random a2;

run;

[SteveDenham]

This should work, and give estimates that are conditional on the random effects.

Steve Denham

[coollord]

I actually run it like this and it says R-side random effects now allowed:

Proc glimmix data = blah plots=residualpanel;

Class a1 a2 a3;

Model ff = a1/ dist = poisson  s;

random a2/ subject = a3 residual type = cs;;

random a2;

run;