I need to estimate log-linear models using sampling weights. I am
interested only in finding the best fitting models, not in parameter
estimates.
My understanding is that this can be done by using the - desmat: poisson
- command to estimate the unweighted cell counts while using the -
exposure - subcommand used to specify the weights. Is this correct? If
so, what goes in the parentheses after - exposure -? I am guessing that
it must be a variable equal to something like the weighted count divided
by the unweighted count.
Any insight is greatly appreciated.
Paul
--
Paul Perl
Center for Applied Research in the Apostolate (CARA)
2300 Wisconsin Ave. NW, Suite 400
Washington, DC 20007
Direct phone (preferred): (202) 687-1365
General CARA phone: (202) 687-8080
Fax: (202) 687-8083
[email protected]
http://cara.georgetown.edu
*
* 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/
