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]

Re: st: propensity score matching with caliper


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: propensity score matching with caliper
Date   Wed, 24 Feb 2010 16:04:00 -0500

Scott Strassels <scotts1@mail.utexas.edu> :
Maybe you just don't have any close matches?  Have you looked at the
distribution of propensity scores by treatment status?

sysuse nlsw88, clear
qui logit collgrad south smsa c_city industry occupation union married
never_married
predict p
set seed 1000
g x = uniform()
sort x
psmatch2 collgrad, pscore(p) neighbor(1) ate noreplace caliper(0.15)
sort _id
gen match = id[_n1]
gen treat = id if _nn == 1
li match treat id _id _n1 in 1/10
sysuse nlsw88, clear
qui logit collgrad south smsa c_city industry occupation union married
never_married
predict p
set seed 1000
g x = uniform()
sort x
psmatch2 collgrad, pscore(p) neighbor(1) ate noreplace caliper(0.015)
sort _id
gen match = id[_n1]
gen treat = id if _nn == 1
li match treat id _id _n1 in 1/10
sysuse nlsw88, clear
qui logit collgrad south smsa c_city industry occupation union married
never_married
predict p
set seed 1000
g x = uniform()
sort x
psmatch2 collgrad, pscore(p) neighbor(1) ate noreplace caliper (0.001)
sort _id
gen match = id[_n1]
gen treat = id if _nn == 1
li match treat id _id _n1 in 1/10


On Wed, Feb 24, 2010 at 3:29 PM, Scott Strassels
<scotts1@mail.utexas.edu> wrote:
> Hello everyone,
>
> I have a question about propensity score matching using psmatch2 with Stata 11.  I've been digging through the literature and Statalist archives on the topic without much luck.  I think I'm using psmatch2 like other authors have, but I don't understand the results.
>
> My colleagues and I have already generated the propensity score.  Here's what I've been doing:
>
> set seed 1000
>
> . gen x = uniform()
>
> . sort x
>
> . psmatch2 treatm, pscore(propen) neighbor(1) ate noreplace caliper (0.15)
>
> . sort _id
>
> . gen match = id[_n1]
> (309 missing values generated)
>
>
> . gen treat = id if _nn == 1
> (309 missing values generated)
>
> . drop if treat == .
> (309 observations deleted)
>
> . save "/Users/scott/Desktop/propensity/Pali vs Ris caliper 022310.dta", replace
> (note: dataset contains 0 observations)
>
> Any suggestions would be very much appreciated.
>
> Thanks,
>
> Scott

*
*   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