Dear all,
I am analysing three datasets, each of which
are from the same population bu different years,
say 2002, 2003 and 2004, but not longitudinally
coded;
The number of cases of a disease (numb_of_cases),
the population size (popul)
the degree of exposure (exposure)
are recorded in each of 1000 towns.
The incidence (number of cases per
population) is about 1/1000 per year.
Initially, I used such commands as
.glm numb_of_cases exposure if year==2002, f(b popul)
.glm numb_of_cases exposure if year==2003, f(b popul)
.glm numb_of_cases exposure if year==2004, f(b popul)
Q1. I would like to bundle these three years
into one analysis. "cluster" option is
available for such a command as "regress",
but not for "glm". How could this analysis
be clustered?
Q2. I was suggested that
familly binomial may not be adequate;
logistic or poisson regression may be more
suitable. What commands can be used for
the above settings?
Let me apologise to you for these lay
questions.
Nagao
__________________________________
Save the earth
http://pr.mail.yahoo.co.jp/ondanka/
*
* 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/
