st: saving case ids when using "propensity score" matching

["Weichle, Thomas" <Thomas.Weichle@va.gov>]

Hi Ben,
I am using -psmatch2- by Leuven and Sianesi.  Here is how I have been
creating match IDs for 1-1 nearest neighbor matching using the variables
that are produced from -psmatch2-.

* Creating common match ID among each pair of treated and controls
For without replacement:
generate _matchid = _id if _weight == 1
replace _matchid = _n1 if _treated == 1

For with replacement:
generate _matchid = _id if _weight != .
replace _matchid = _n1 if _treated == 1

The control cases that are being used more than once in a "with
replacement" match can be accounted for in an outcome analysis by using
frequency weights ([fweight = _weight]).  Elizabeth Stuart provides a
nice review of matching method concepts and weighting issues in a
manuscript titled "Matching methods for causal inference: A review and a
look forward" (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2943670/).

Unfortunately, I haven't been able to figure out how to create a common
match ID for k-1 nearest neighbor matching (k>1) which restricts the
matching to "with replacement".

Tom Weichle
Math Statistician
VA Information Resource Center (VIReC)
Edward Hines Jr. VA Hospital, Bldg 18, 202E
708-202-8387 ext. 24261
Thomas.Weichle@va.gov


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/