Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: logit / odds ratio interpretation using -listcoef-


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: logit / odds ratio interpretation using -listcoef-
Date   Sat, 15 Mar 2008 15:26:43 +0000 (GMT)

Yes, that can happen, for instance in the example below the odds of
being in good repair is ~1900% larger for foreign cars than for
domestic cars: 

sysuse auto, clear
recode rep78 1/3=0 4/5=1
logit rep78 foreign , or
listcoef, percent

In this case the odds ratios compares the likelihood of being in good
repair of foreign and domestic cars. The likelihood is represented with
odds: the number of carrs in good repair for every car in bad repair.
To gain a bit more confidence in this result you might want to know
what the odds of being in good repair is for domestic cars. This you
would get if Stata also reported exp(_cons) when you asked for the -or-
option, unfortunately Stata doesn't do that (I know, I have asked for
that a number of times). However, you can trick Stata in doing that:

gen one = 1
logit rep78 one foreign , or nocons

For every domestic car in bad repair there are only ~.30 cars in good
repair. If we multiply this by the odds ratio we get the odds of being
in good repair for foreign cars:

di exp(_b[foreign])*exp(_b[one])

For every foreign car in bad repair there are 6 cars in good repair. We
could get those two odds in one go by adding both a dummie for domestic
and a dummy for foreign cars (and leave out the constant):

gen domestic = !foreign
logit rep78 domestic foreign, nocons or

I hope this helps,
Maarten


--- Jn <ensam21@gmail.com> wrote:

> Dear Stata users,
> 
> I am a student and I have one question regarding interpretation of a
> logit regression. After running my regression I used -listcoef-
> command (in percentage form) for interpretation of the coefficients.
> Some of my dummy independent variables return with extremely high
> percentage chnage in odds i.e. 2,500% and I am not sure if I am
> interpreting this right. Does that mean the odds of my dependent
> variable = 1 increase by 2500% for ___??___ (i don't get this part,
> when the independent variable is binary). Would it make sense for me
> to interpret it as, for instsance, a 0.5 increase in probability of x
> variable increases the odds of y=1 by 1250 %?
> 
> Any help appreciated..thanks
> *
> *   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/
> 


-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

http://home.fsw.vu.nl/m.buis/
-----------------------------------------


      ___________________________________________________________ 
Rise to the challenge for Sport Relief with Yahoo! For Good  

http://uk.promotions.yahoo.com/forgood/
*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index