Bioequivalence and Bioavailability Forum

Dear all & esp. Detlew!

I know that Balaam’s design (RT|TR|RR|TT) is not very powerful since only ½ of the sequences are “used” in the T-R comparison (the other half is luxury – giving s²_WR and s²_WT). What puzzles me is

require(PowerTOST)
sampleN.TOST(CV=0.3, design="2x4x2", print=F)
  Design alpha  CV theta0 theta1 theta2 Sample size Achieved power Target power
1  2x4x2  0.05 0.3   0.95    0.8   1.25         152      0.8067485          0.8

compared to

sampleN.TOST(CV=0.3, design="2x2", print=F)
  Design alpha  CV theta0 theta1 theta2 Sample size Achieved power Target power
1    2x2  0.05 0.3   0.95    0.8   1.25          40      0.8158453          0.8

152/2=76 subjects in Balaam’s design resemble TR|TR which is obviously ≫ than the 40 needed in a conventional 2×2.
If I politely ask

power.TOST(CV=0.3, design="2x2", n=76)
[1] 0.9732043