HIGHLIGHTS
SUMMARY
Sadly this in not the first time-or the last time-that such staggering quality disasters are revealed. If the stated criterion is to control Type-II error first, e_g, the test must be at least 80% powerful under a specified alternative hypothesis, then indeed the authors may have to find a (non-trivial) lower bound of the variance or standard error (such as SE(θ̂1 + θ̂2 ) ≥ |SE(θ̂1 ) - SE(θ̂2 )| when SE(θ̂1 )=SE(θ̂2 )). From an inferential point perspective, this increased confidence should be welcomed not only because it helps to guard . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.