Date: Nov 6, 2012 6:17 PM
Author: Luis A. Afonso
Subject: The null is almost surely false . . .
The null is almost surely false . . .
Intuition is repeatedly used everyday life. An example: when we wish to know what someone exactly want to say, but not wanting to be clear, because of shyness, discretion, not be heard by the surrounding people, the presence of a police officer, or whatever.
A statement currently believed by the anti-NHST people is: The Null Hypotheses, H0, is almost surely false; to be concerned if is true or not true, worthless.
Using intuition does uncover two errors the guy thinks about NHST.
Firstly he thinks that the objective is to decide the H0 veracity. And read (testing a parameter) H0: t=0 as it was an algebraic equation, so the t nullity.
Quite simple to answer properly:
__a) The Significance tests have only one purpose; to reject or not the Null,
__b) the pseudo equation t = 0 means a very different thing: is data so imprecise that t could not be say the test statistics (modeling t) is different from 0? Or not?
Luis A. Afonso