Date: Tue, 14 Oct 2003 13:49:55 -0400
Reply-To: "Burleson,Joseph A." <burleson@up.uchc.edu>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: "Burleson,Joseph A." <burleson@up.uchc.edu>
Subject: Re: logistic regressin power calculations
Content-Type: text/plain; charset="us-ascii"
Gene: Re: comment on Cohen: Cohen's power book will explicitly NOT deal
with repeated-measures of a dichotomous, it will not even deal with
multiple predictors of a dichotomous (logistic regression). I found the
hard way that PASS will not do it, either. My next suggestion is a
conditional logistic regression program, OSIRIS, I believe? That you may
be able to fool into giving you a power by creating some data of known
effect sizes. The other thing you might look into is MPLUS by Muthen &
Muthen, although it is not a power calculation program, it covers
complex designs like repeated measures on a dichotomy. My math/stat
colleagues inform me that there is not any procedure for even analyzing
repeated measures on a dichotomy that they think is valid, to begin
with!
-Joe Burleson
-----Original Message-----
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
emaguin@acsu.buffalo.edu
Sent: Monday, October 13, 2003 4:40 PM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: logistic regressin power calculations
All,
I'm looking for a program that does power calculations for two group
pre-post design with a dichotomous dependent variable. So the analysis
model is the multiple logistic regression model with the time 1 DV as a
covariate and group as the IV predicting time 2 DV. I have talked to
Jerry
Hintze at NCSS/PASS and he says his program won't do it. So i'm
wondering
what else is out there that might do this. Or, does someone know the
citations to the method.
Thanks, Gene Maguin
| 
