Subsampling

Im trying to do a subsampling problem(unbalanced subsamples) with two fixed
treatments. I know how to use glm, and the appropriate ANOVA table/error
setup for the single treatment case. While I have suspicions on what to do,
ANOVA/error term wise, for the two treatment case Im wondering if I could
just simplify life and use proc mixed. Would something like

proc mixed;
class trta trtb id;
model y=trta*trtb;
repeated / type=whatever subject=id;
run;

be reasonably justifiable?

Thanks in advance,
--Scott


.

Quantcast