Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Testing predictor coefficients in ordinal logistic regression


From   jaswartz <[email protected]>
To   [email protected]
Subject   st: Testing predictor coefficients in ordinal logistic regression
Date   Sun, 11 Jan 2004 17:07:51 -0600

All,

As I understand ordered logistic regression, the predictor coefficients (and 
odds ratios) are based on the assumption that the pedictor variable has the 
same effect (slope) across all categories of the dependendent variable. In the 
models I am working with, I do not believe this is a tenable assumption. In 
fact, one of my hypotheses is that serious mental illness has a larger effect 
in moving a person from drug use to dependence on that drug (categories 2 and 
3 for the dependent variable - drug use/dependence) than it does on moving 
people from no use to use (categories 1 and 2). For this reason, I have been 
using multinomial logistic regression and, in fact, the odds ratios for 
serious mental illness and use-dependence are greater than the odds ratio for 
SMI and no use - use. My question, after that longish preamble, is can you 
compare the odds ratios against each other to see if they are significantly 
different? Or equivalently and perhaps more understandably, can you test the 
ordered logistic regression assumption that the effect size for a predictor 
variable is the same across the ordered levels of the dependent variable?

Thanks for any help!

James 
[email protected]

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index