At 12:20 PM 12/11/2002 -0800, you wrote:
Hello,
We've been using the xtnbreg command to look at some crime data. Our unit
of analysis is city*day. Some cities are much larger than others, however,
so we would like to perform a specification in which we weight the cities by
population. We would typically do this with the pweight option, but this is
unavailable. iweights are possible but the standard errors are grossly
incorrect. What can be done to compute the correct standard errors when
using the iweight option?
I guess I wouldn't think of this as a weighting issue so much as an
exposure issue. You have an incidence rate of crime that you are trying to
model, but you observe counts of crimes, which depend on the size of the
population that is exposed to the incidence rate. See for example the
discussion of the exposure option in the manual discussion of _poisson_.
//Jesper
Thanks,
Lars
Lars Lefgren
Assistant Professor
Economics Department
Brigham Young University
130 FOB
Provo, UT 84602-2363
(801) 422-5169
Lars Lefgren
Assistant Professor
Economics Department
Brigham Young University
130 FOB
Provo, UT 84602-2363
(801) 422-5169
*
* 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/
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Jesper B. S�rensen
Richard S. Leghorn (1939) Associate Professor of Strategic Management
Sloan School of Management
Massachusetts Institute of Technology
E52-581
Cambridge, MA 02142
http://web.mit.edu/sorensen/www/
(617) 253 7945 -- voice
(617) 253 2660 -- fax
*
* 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/
