Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Re: Estimating a model where the dependent variale is a ratio


From   "Joseph Coveney" <jcoveney@bigplanet.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Re: Estimating a model where the dependent variale is a ratio
Date   Fri, 4 Sep 2009 09:45:11 +0900

Johannes Schoder wrote:

I am struggling for a while with the following issue:
I should estimate a  model of the following form:
(Number of people diagnosed with cancer type i in county c who survived 
5 years or more)/ (Total number of people diagnosed with cancer type  i 
in county c)=beta_ln_Number of people diagnosed with cancer type i in 
county c

In my dataset I have information on the survival time in months for each 
individual, the county where he was diagnosed, and the cancer type he 
was diagnosed with.
My advisor is using SAS and used the command:
 PROC GENMOD; class type county MODEL N_SURV/N_DIAG=ln_DIAG type 
REPEATED SUBJECT=county
However, in Stata I don't know how to estimate a model where the 
dependent variable is a ratio (see the lhs variable above). Does anyone 
know how to estimate this kind of model in Stata?

--------------------------------------------------------------------------------




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index