Re: Comparison of Logistic regression and Cox regression
- From: Richard Ulrich <Rich.Ulrich@xxxxxxxxxxx>
- Date: Sun, 10 Jul 2005 20:39:03 -0400
On 7 Jul 2005 03:48:16 -0700, ramsathish@xxxxxxxxx wrote:
> Respected Sir,
>
> I am Ramsathish S doing M.Phil (STATISTICS) in Loyola
> College, Chennai, India. I have one doubt sir, How to compare the odds
> ratio from logistic regression and the hazard ratio from the cox
> regression.
Logistic regression has a single dichotomous outcome.
Cox regression models a survivorship curve.
The data are different. What other comparison do you need?
I am curious if anyone can describe the difference between
the Cox regression hazard ratio and the Mantel-Haenszel (1959)
odds ratio (called at the time, relative risk).
- MH computes across 2x2 tables. Do I remember right, that
there is a version valid for survivorship data across time?
- The MH statistic uses Pearson-type Expected-Observed for
computing a test. I think the Cox procedure is maximum
likelihood: Is that so? Is there any other difference?
--
Rich Ulrich, wpilib@xxxxxxxx
http://www.pitt.edu/~wpilib/index.html
.
- Follow-Ups:
- Re: Comparison of Logistic regression and Cox regression
- From: Kevin E. Thorpe
- Re: Comparison of Logistic regression and Cox regression
- References:
- Comparison of Logistic regression and Cox regression
- From: ramsathish
- Comparison of Logistic regression and Cox regression
- Prev by Date: How to Begin?
- Next by Date: extreme value
- Previous by thread: Comparison of Logistic regression and Cox regression
- Next by thread: Re: Comparison of Logistic regression and Cox regression
- Index(es):
Relevant Pages
|
