Date: Oct 28, 2012 5:07 PM
Author: Luis A. Afonso
Subject: p-value < alpha, correct to reject the Null!
Felt in the rejection interval, Firr
A few people angrily claim, that the p-values commonly obtained in NHST are NOT, is even far away, what definition states:
(From Wikipedia, the free encyclopedia): in statistical hypothesis testing, the p-value is the probability of obtaining a test statistic at least as extreme as the one that was actually observed, assuming that the null hypothesis is true.
In fact, what all people do is to obtain a value that is simply p-value = 1 - F(z0), from F(z) , the Test Statistics Cumulative Distribution. Then if p-value < alpha (a significance level chosen previously) we decide that the null hypotheses H0 is very unlike to be true; therefore we reject H0 in favour of Ha, the alternative one.
Why do not change to p-value to *firr* and keep the procedure: with the commonly chosen are only a 1/20 probability the *test* fall into the rejection interval if H0 is true. If it does we can be pratically sure that the null is false.
(Down the p-value, up the firr!)
Luis A. Afonso