Monica wrote:
-------------------------------------------------
I am are using "matched" data (nearest-neighbor matching based on
propensity scores) and thus the standard errors (and test statistics)
for Stata's oprobit coefficients are incorrect because oprobit's
variance formula ignores the additional variance that arises from
multiple matches to the same unit (i.e., the matching process itself
adds variability to the sample that must be incorporated into the
var-cov estimate). I therefore use a variance formula supplied by
Abadie and Imbens (2004), but I do not know how to incorporate this
adjusted variance estimate for the sample into an ordered probit
calculation. If I were doing OLS, I would just calculate the standard
errors by inserting the adjusted variance estimate into the standard
formula, but I am unsure how I can use our variance estimate to generate
the standard errors of the ordered probit coefficient estimators.
-------------------------------------------------
If the matched sets are identified by the variable setid, wouldn't the
-cluster(setid)- option do what you want?
Svend
__________________________________________
Svend Juul
Institut for Folkesundhed, Afdeling for Epidemiologi
(Institute of Public Health, Department of Epidemiology)
Vennelyst Boulevard 6
DK-8000 Aarhus C, Denmark
Phone: +45 8942 6090
Home: +45 8693 7796
Email: [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/