## A first look [BE/BA News]

Hi all!

❝ ’In the Random statement, TYPE=FA0(2) could possibly be replaced by TYPE=CSH or UNR.’ THX for CSH. Why not TYPE=FA0(1)?

❝ Instead of Satterthwaite’s degrees of freedom Kenward-Roger could be used (what does ‘possibly’ mean here?)

Well it finally happened. It took 20 years to find out that FA0(2), CSH and UN is the same for this model for 2 formulations (but possibly means that not entirely sured about this). May be possibly means that in is not true for 3 formulation?

I don't understand what FA0(1) means for this model? No correlation? Same as diagonal cov structure?

❝ Alternative software could also be used if same results are generated as in PROC MIXED in SAS.

It is very funny. Because if percept it exactly - there is no software that give you exactly same DDoF (Satterthwaite’s or freedom Kenward-Roger). All of them (SPSS, WinNonlin ets...) give you slightly different results for DDoF.