Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

st: pscore vs. psmatch2


From   Yee Kyoung Kim <icejean@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: pscore vs. psmatch2
Date   Thu, 2 Dec 2010 00:48:51 +0900

Dear Statalists,

I've been trying to calculate ATT using both pscore and psmatch2.

(1)
pscore Treat x1 x2 x3 , pscore(mypscore) blockid(myblock) comsup
numblo(5) level(0.005) logit
set seed 1221
attnd y Treat x1 x2 x3  , pscore(mypscore_ml) comsup boot reps(100) dots logit


(2)
psmatch2 Treat x1 x2 x3, outcome(y) logit

I confirmed that propensity scores are exactly the same in (1) and (2).

Q1: If I run psmatch2, I can get both unmatched and matched ATT as shown below.
       Then what about "pscore"?  Is ATT in Table 2 below based on the
matched sample?

<Table 1: psmatch2 results>
----------------------------------------------------------------------------------------
        Variable     Sample |    Treated     Controls   Difference
    S.E.   T-stat
----------------------------+-----------------------------------------------------------
         y  Unmatched | 2.73645662   1.34825899   1.38819763
.24860504     5.58
                        ATT | 2.73645662   1.44619683   1.29025979
.954900678     1.35
----------------------------+-----------------------------------------------------------

<Table 2: pscore results>
ATT estimation with Nearest Neighbor Matching method
(random draw version)
Analytical standard errors

---------------------------------------------------------
n. treat.   n. contr.         ATT   Std. Err.           t
---------------------------------------------------------

       85          27       0.077       0.073       1.061

---------------------------------------------------------
Note: the numbers of treated and controls refer to actual
nearest neighbour matches


Q2: How can I get the same results in both (1) and (2). In order to
get the result of (1), how can I revise the command of (2)
Actually, I tried with "common" option in (2) as below, but the
results are quite more different.


(2)' psmatch2 Treat x1 x2 x3, outcome(y) logit common



Thank you so much, as always.

Best,
YK
-- 
Yee Kyoung Kim
Korea Institute of Intellectual Property
Associate Research Fellow
010-2595-6148
icejean@gmail.com
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index