Re: A simple QUIZ

Jack Tomsky wrote:
A simple QUIZ



The candidates are only those that are persuaded that
the null hypotheses could be proved TRUE.(for example
Jack Tomsky, OMU, Bob Ling – Reef Fish, David
Winsemius).


*** How many flips should I perform in order to prove
that a coin is fair (probability of heads up = ½)
***
(I can produce a full hand of examples)

*******
licas


Afonso has never understood that the role of statistical inference in hypothesis testing is to make decisions in the face of uncertainty. That's why we have type I and type II errors.

Under the Afonso theory of hypothesis testing, the only permissible decisions are either to reject the null hypothesis or to make no decision. Thus, the power function is always zero.

Jack

Jack, I think you may be mis-characterizing what Luis is trying to say. I believe his position is the same one taken by Altman & Bland in their article, "Absence of evidence is not evidence of absence". It can be downloaded at the link given below. Here is the opening paragraph:

"The non-equivalence of statistical significance and clinical importance has long been recognised, but this error of interpretation remains common. Although a significant result in a large study may sometimes not be clinically important, a far greater problem arises from misinterpretation of non-significant findings. By convention a P value greater than 5% (P>0.05) is called "not significant." Randomised controlled clinical trials that do not show a significant difference between the treatments being compared are often called "negative." This term wrongly implies that the study has shown that there is no difference, whereas usually all that has been shown is an absence of evidence of a difference. These are quite different statements."

http://www.bmj.com/cgi/content/full/311/7003/485?q=y

From this point of view, the two decisions are "reject H0" and "fail to reject H0", and for the reason given by Altman & Bland, the latter is not seen as equivalent to "accept H0"--except perhaps in the case of (bio-)equivalence testing. But in that case, one would ensure that there was adequate power to detect a difference of a given (clinically important) size.

You have cited some examples of well-known statisticians who were happy to use "the accept H0" terminology. It would not surprise me if some equally eminent statisticians have eschewed that terminology. Does anyone have any examples of this?

Cheers,
Bruce

--
Bruce Weaver
bweaver@xxxxxxxxxxxx
www.angelfire.com/wv/bwhomedir
"When all else fails, RTFM."
.

Relevant Pages

  • Re: A simple QUIZ
    ... Under the Afonso theory of hypothesis testing, ... "Absence of evidence is not evidence ... statisticians who were happy ...
    (sci.stat.math)
  • Re: Sally Clark died of alcohol poisoning
    ... evidence that he, quite frankly, did not really understand. ... It is axiomatic among statisticians that very few - if any - medical ... was giving evidence on the alleged likelihood of recurrent "cot death" ...
    (uk.legal)
  • Re: A simple QUIZ
    ... evidence of a difference. ... and for the reason given by Altman & ... Bland, the latter is ... statisticians who were happy ...
    (sci.stat.math)
  • Re: A * Titans * Fight III : Jack Tomky versus R A. Fisher
    ... WHAT Jack Tomsky thinks is directed to? ... What I said is the current way true Statisticians use to draw conclusions from Hypotheses Tests. ... Luis Amaral Afonso [The moderator destroyer] ...
    (sci.stat.math)
  • Re: The Formula for D
    ... The most one can say is that there was no sufficient evidence to reject H0. ... (Jack Tomsky in view of the tremendous ignorance he is shown in various Statistical matters is not allowed to call Statisticians to whatever: he unquestionably proves that do not belongs to this set of persons). ...
    (sci.stat.math)
Quantcast