Dear sci.stat.edu readers,
This may be a little late but if you would like an electronic reprint
(PDF) of that short post hoc power paper by Zumbo and Hubley that Rich
Ulrich mentioned earlier in this thread, feel free to send me an e-mail
message at [EMAIL PROTECTED] and I will send one along via e-mail right
away.
Please do not respond to this e-mail message in requesting a
reprint. Instead please send me an e-mail message directly.
Cheers, Bruno
Bruno D. Zumbo, Ph.D.
The University of British Columbia
2125 Main Mall, Scarfe Building, Dept. of ECPS
(Program area: Measurement, Evaluation, and Research Methodology)
Vancouver, B.C. V6T 1Z4 CANADA
e-mail: [EMAIL PROTECTED]
webpage: http://www.educ.ubc.ca/faculty/zumbo/
---------- Forwarded message ----------
(beginning of original message)
Subject: Re: Post hoc power
From: [EMAIL PROTECTED]
Date: 2000/10/06
Newsgroups: sci.stat.edu
( better late than never. )
This might be relevant to Steve's request.
On 1 Sep 2000 16:20:13 -0700, [EMAIL PROTECTED] (Simon, Steve, PhD)
wrote:
> I need to find a good reference that critiques the use of post hoc power
> calculations--in particular those calculations that use the observed effect
> size rather than an effect size which is clinically relevant.
>
"A note on misconceptions concerning prospective and retrospective
power," by Bruno D. Zumbo and Anita M. Hubley. The Statistician
(1998) 47, Part 2, pp. 385-388.
... ' We suggest that it is nonsensical to make power calculations
*after* a study has been conducted....' And they say, report the
effect size.
They say, "retrospective power" only makes sense in a (non-research)
setting where you know the ultimate truth and falsity.
http://www.pitt.edu/~wpilib/index.html
(end of original message)
------------------------------------------------------------------------------
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
