Re: The Logic of Hypotheses Tests

I made “thousands “ of Hypotheses Tests in my entire
life and not even one time I made confusion (surely
because I use the version CRITICAL VALUES): therefore
the concept NOT ACCEPT – FAILED TO ACCEPT is
absolutely clear in my mind.
And so clear that, when I passed to the, not usual,
version p(data| H0) I didn’t realize that the
condition leading to FAIL TO REJECT (NOT SAY TO
ACCEPT!!!) is, in fact,
____________ ALPHA > 1 – p(data| H0)

My humble apology. Conclusion and advertising: use
preferentially the concept of critical values!


Luis Amaral Afonso



In Afonso 2.0, the decision rule is now

do not accept H0 at the ALPHA significance level (usually 5% or 1%) only if:

_____p(data | H0) < ALPHA

and FAIL TO REJECT (NOT SAY TO
ACCEPT!!!) if:

____________ p(data | H0) > 1 - ALPHA.


Thus, when

ALPHA <= p(data | H0) <= 1 - ALPHA,

the decision rule is undefined.

Because of these poor characteristics, I recommend that neither Afonso 2.0 nor Afonso 1.0 be used in hypothesis testing as replacements for Neyman-Pearson.

Jack (moderator)
.