Re: Gam Address & Healthy survivor effect -Reply -Reply

["Fritz A. Seiler" <faseiler@nmia.com>]

Mark Hanlon wrote:

> Professor Cohen's comment is quite true, of course, but I can't help
> remembering a time when I went to a Biostatician with some data, and he
> asked : "well, what would you LIKE me to prove ?".

That biostatistitian was definitely the wrong person to ask!  You CANNOT PROVE
ANYTHING by a statistical test. All you can do is one of two things: Either you
can reject the null hypothesis at a 1 - alpha confidence level or then you
cannot.  These negative statements are all that statistical testing can do, no
more!

> One must, as Prof. Choen does, clearly understand how data has been
> analysed, and the meaning of any statistics thus generated.

One cannot overstate the importance of Bernie Cohen's remark.

Sorry, but I could not resist.

Best regards

Fritz

*************************

Fritz A. Seiler, Ph.D.
Principal
Sigma Five Associates
P.O. Box 14006
Albuquerque, NM 87191-4006
Tel. 505-323-7848
Fax. 505-293-3911
e-mail: faseiler@nmia.com

**************************