Despite what is common in applied work, if you are going to bootstrap the standard errors, you should stick with kernel matching. Work by Imbens and Abadie have shown that there is no basis for doing bootstrapping with nearest neighbor matching (and I presume the same may also apply to radius matching) due to the non-smoothness of the estimator.
Dann
****************************************************
Daniel L. Millimet, Associate Professor
Department of Economics
Box 0496
SMU
Dallas, TX USA
phone: 214.768.3269
fax: 214.768.1821
web: http://faculty.smu.edu/millimet
****************************************************
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Patricia Vera Rojas
Sent: Tuesday, November 01, 2005 9:57 PM
To: [email protected]
Subject: st: bootstrap after psmatch2
Hi all,
I am deciding which method of matching is the best for my
data (nearest neighbor, k-neighbor, kernel, radius). After
psmatch2 is it correct to use bootstrap command to take one
of them as the best one?
I mean, smallest bootstrapped standard errors could be a
good criteria?. Do them have to be less than one?
Thanks,
Patricia
______________________________________________
Renovamos el Correo Yahoo!
Nuevos servicios, m�s seguridad
http://correo.yahoo.es
*
* 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/
*
* 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/
