Bioequivalence and Bioavailability Forum

Very well, d_labes.

On basis of the current evidence the jury then votes in favour of power approximations as they can be used without any significant error in the power regions of practical interest. Power approximations, including those based on the non-central t distribution, will therefore be considered resurrected and fully viable. Amen. More generally, the jury adds that success in real life depends on a lot of approximations. Examples of other approximations that are fully justified and steer the user free of potential trouble include, but are not limited to:

1. "No honey, you do not look fat in that dress."
   
2. "Yes honey, of course I remembered to invite your aunt Betty for the christmas dinner."
   
3. "Subject 73 can be removed from the dataset because of a protocol violation in period 1".

Best regards
EM.