Reply from ElMaestro, 2012-11-06 11:48:
Dear Helmut,
» Sequence+Stage+Period(Stage)+Treatment+Sequence*Stage+Subject(Sequence*Stage)
»
» Now I get the same results (PE, CI, Means, SEs, :blahblah:) as in PHX’ mixed model, SAS, and R. Don’t run the silly-o-meter on this setup.
Looks like this would kick the the silly-o-meter directly into a cosmic overdrive. I would intuitively say you should get the same results if you remove the Sequence (first) term. Can check?
By the way: If models need to be written anything like that then I would guess period would need to be specified as period(sequence) in the standard 2,2,2-BE model. Just trying to follow the logic that I don't understand. Nicht?