|
Science Forum Index » Statistics - Education Forum » Needed: medical example of a continuous-variable interactio
Page 1 of 1
|
| Author |
Message |
| Bruce Weaver |
 Posted: Fri Nov 24, 2006 8:07 am |
|
|
|
Guest
|
Can anyone suggest a good medical or health-related example of a linear
regression model that has two continuous predictor variables and their
interaction? A public-domain data set would be useful too, if available.
Thanks,
Bruce
--
Bruce Weaver
bweaver@lakeheadu.ca
www.angelfire.com/wv/bwhomedir |
|
|
| Back to top |
|
| Reef Fish |
| Posted: Fri Nov 24, 2006 10:23 am |
|
|
|
Guest
|
Bruce Weaver wrote:
Quote: Can anyone suggest a good medical or health-related example of a linear
regression model that has two continuous predictor variables and their
interaction? A public-domain data set would be useful too, if available.
That should be the easiest thing to find since any "response surface
methodology" would have the interaction of two or more continuous
variables, even for a quadratic surface.
How many of those are good examples is a different question.
However, I think you should look over a CLASS of such methodolgy
used rather than any particular "medical example".
http://www.washington.edu/students/crscat/biostat.html
The webpage lists many examples of COURSES of that type. An
inquiry to the faculty of those courses might give you the needed
references.
-- Reef Fish Bob. |
|
|
| Back to top |
|
| Reef Fish |
| Posted: Fri Nov 24, 2006 10:29 am |
|
|
|
Guest
|
Bruce Weaver wrote:
Quote: Can anyone suggest a good medical or health-related example of a linear
regression model that has two continuous predictor variables and their
interaction? A public-domain data set would be useful too, if available.
As I was suggesting you look at a CLASS of such problems in
response surface designs, I came across this webpage:
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7703830&dopt=Abstract
It's certainly a medical study of the type you sought; whether they
would
release their data OR if they are doing anything worthwhile, I can't
judge
from the short blurp.
Do some web search with keywords like "medical studies using
response surface designs".
-- Reef Fish Bob. |
|
|
| Back to top |
|
| Bruce Weaver |
| Posted: Tue Nov 28, 2006 5:48 pm |
|
|
|
Guest
|
Bruce Weaver wrote:
Quote:
Thanks Bob, that might help. I didn't think of using "response surface"
as one of the search terms. And David, thanks for your suggestions too.
As it turned out, I had better luck finding examples when I searched on
"moderated regression" and health or medical example. "Moderation" is
the term that <tongue in cheek> RF Bob's old friends Barron & Kenny
</tongue in cheek> use to describe interaction in linear regression
models. ;-)
--
Bruce Weaver
bweaver@lakeheadu.ca
www.angelfire.com/wv/bwhomedir |
|
|
| Back to top |
|
| Reef Fish |
| Posted: Wed Nov 29, 2006 1:18 pm |
|
|
|
Guest
|
Bruce Weaver wrote:
Quote: Bruce Weaver wrote:
Thanks Bob, that might help. I didn't think of using "response surface"
as one of the search terms. And David, thanks for your suggestions too.
As it turned out, I had better luck finding examples when I searched on
"moderated regression" and health or medical example. "Moderation" is
the term that <tongue in cheek> RF Bob's old friends Barron & Kenny
/tongue in cheek> use to describe interaction in linear regression
models. 
The name Kenny is still hurting me cheek from tongue poking and
laughing at the same time, But Barron? Never heard of him or her.
Even if David plays the guitar, I don't think you're referring to Babe
Barron. <turn the other tongue at ya>
http://www.mmguide.musicmatch.com/album/album.cgi?ALBUMID=1388478
Perhaps Kenny gave up his academic career and went into rock
band music and called his band "Moerated Regression", huh?
-- Reef Fish Bob,
|
|
|
| Back to top |
|
| |
