Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Propensity Score Matching - Show matches

From   daniel klein <>
Subject   Re: st: Propensity Score Matching - Show matches
Date   Wed, 2 Nov 2011 11:28:07 +0100


please say where user-written commands come from. You are using
-pscore-, which is probably

package st0026 from

      SJ2-4 st0026.  Estimation of averate treatment effects ...

and its latest update

package st0026_2 from

      SJ5-3 st0026_2.  Estimation of average treatment effects

      Estimation of average treatment effects
      by Sascha O. Becker, University of Munich
         Andrea Ichino, EUI
      After installation, type help attk, attnd, attnw,
       attr, atts, and pscore

As far as I can see, your options are given in the help files. For
example the first sentence in the help file for -pscore- states that

"pscore [...] stratifies individuals in blocks according to the
pscore; displays summary statistics [...] of the stratification;"

So I would start with options -detail- and -blockid-. Also, in the
help file for -attnw- (or -attnd-) (both links are in the help file
for -pscore-) you find other options ( -match_options-  -matchvar-,
-matchdta-, and -id-) that might be helpfull.

I was not able to get exaclty what you want (in a 5 min trial).
However, I am wondering just why you would like to see a list of the
nearest neighbors anyway?

I know it is (or was, or has been) disapproved to make suggestions on
alternative user-written commands, but of you do not intend to stick
to -pscore-, you might want to try -psmatch2- (available from SSC). In
the help-file you will find, that variable "_nk" is created.

"_nk In the case of one-to-one and nearest-neighbors matching, for
every treatment observation, it stores the observation number of the
k-th matched control observation."


Dear all,
I am doing Propensity Score Matching with pscore and I would like to
know if there is a way that Stata can give me a table (or put it in a
variable with IDs or something similar) that shows the matches, so for
example it shows the treated object on the left and the 5 nearest
neighbors (of the control group) on the right? Or is there any other
way of assigning the 5/10/... nearest neighbors of the control group
to each treated object?
Thanks, Maximilian
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index