Radford Neal wrote: > Jeffrey's > criticism obviously applies just as much to people doing fixed level > testing.
I would argue not as much. I agree fixed level testing does rely on the small probability of things not seen, but the criticism is largely blunted or made irrelevant because the rejection rule is selected prior to examining the data. If P values are not described in terms of fixed level testing (as is true of many explanations), then I don't know how to defend them. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
