Date: May 22, 2013 8:17 PM
Author: Luis A. Afonso
Subject: 1- NHST
[Preventing abusive practice, critics and claims to ban NHST´s from scientific literature]
I do not need, sure, to add further reasons than the following one to agree that the condition pvalue < = alpha strongly suggests that the Null Hypotheses, H0, is false. In fact one is enough: if so all p-values do follow a Uniform Distribution in [0, 1], so alpha, ordinary set 0.05, is the probability that the statement H0=true is indeed false.
After presetting alpha and performing the test, if:
____p-value > alpha, we did not get sufficient evidence to reject the Null Hypotheses, H0,
____p-value <= alpha we reject H0 with the probability 1-alpha confidence, therefore alpha level of significance.
Luis A. Afonso