Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: NHST
Replies: 1   Last Post: Oct 12, 2012 9:05 PM

 Messages: [ Previous | Next ]
 Luis A. Afonso Posts: 4,758 From: LIsbon (Portugal) Registered: 2/16/05
NHST
Posted: Oct 11, 2012 11:20 PM

NHST

Tough Science was moved by elective praxis Null Hypotheses Significance Tests (NHST) was yet absolutely discarded facing the frequency they are attacked and violence as well. With the periodicity of monsoons or insect plagues we have to hear a *lucid* fellow more, to inform the ignorant tribe (all of us) of two important things: 1) The null hypotheses is always false, 2) The test, if not rejected, (they say accepted, of course), does not tell whatever about the difference between the observed parameter value and the reference one.
The question crux is that they do not understood that they use to find tests of ZERO: the rejection meaning that the null hypotheses cannot be true (at the set significance level) which is an wrong practice and not a test, only slightly modified and present at all elementary text-books, envisaging to detect if a convenient to the research purpose DIFFERENCE can be warrant. Add to the test formula a subtractive (or additive) quantity . . . *et voilà *.
About the impossibility the Null be true, is so stupid that it no deserves a comment: as H0=0 did not signify another completely different thing: that we are unable (given data, and significance level) to get sufficient evidence that the nullity is false.

The last *bright* scientist I found,
Esa Laara, Statistics: reasoning on uncertainty and the insignificance of testing null.
See: