folks --
i have a sample of students who have pweights (inverse of the
probability of selection into the sample). my outcome variable is a
test score, and for each student, i have an estimate of the
person-specific standard error of their test score, the inverse square
of which is an analytic weight (it represents the inverse of the error
variance of the test score).
my question is, how should i combine these weights in a regression
model?
my (intuitive) thought is to multiply them, and then to tell Stata that
the resulting product is an aweight -- can anyone let me if that
intuition makes sense, or if there is a more appropriate approach?
specifically, if i have the variables
sprob (sampling selection probability)
setest (standard error of test score)
test (outcome test score)
then would this be appropriate:
g pwt=1/sprob
g ewt=1/setest^2
g combwt=pwt*ewt
reg test varlist [aw=combwt]
thanks for any help, advice.
best,
sean.
_____________________________
sean f. reardon
associate professor of education and (by courtesy) sociology
school of education
485 lasuen mall, #315
stanford university
stanford, ca 94305-3096
650.736.8517 (office phone)
650.725.7412 (office fax)
[email protected]
_____________________________
*
* 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/
