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 at the end of May, and its replacement, statalist.org is already up and running.


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

re: st: Statistical Matching with psmatch2 ?


From   "Ariel Linden, DrPH" <ariel.linden@gmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   re: st: Statistical Matching with psmatch2 ?
Date   Fri, 31 May 2013 09:51:42 -0400

It is not clear what you have done here. It sounds like you have two
different datasets and you ran these programs separately in each data set?
That doesn't make much sense. You should have all the data in one file,
estimate the propensity score on the entire pool, do your
matching/weighting, and then run the outcome model.

It is also not clear how or why you have missing data? Are these data
missing at random? Is there substantial data missing? Would simple or
multiple imputation be sufficient?

These are questions you need to think about...

Ariel

Date: Thu, 30 May 2013 19:48:15 +0200
From: Christian Miranda Pacheco <Christian.Miranda@univ-paris1.fr>
Subject: st: Statistical Matching with psmatch2 ?

Hi everybody,

I try  to match two different datasets A(contaning variables a)and B  
(containing variables b). using propensity score. They have common  
demographic covariates c(age, sex..) and i would like to use them to  
get a unique dataset, from this form


id    (c)    (a)    (b)
1A     c      a      .
2A     c      a      .
3A     c      a      .
4B     c      .      b
5B     c      .      b
6B     c      .      b

To the merged dataset (a') I try to get _b being the estimated  
variables taken from (b) , T being the tratment status

id    T    (c)    (a)    (b)    (a')
1A    0     c      a      .      a
2A    0     c      a      .      a
3A    0     c      a      .      a
4B    1     c      .      b      _b
5B    1     c      .      b      _b
6B    1     c      .      b      _b

I'am actually using .pscore and .psmatch2
First i get a balanced pscore from,

.pscore T c ,pscore(PSCORE)

Then I include it in the following precedure

.psmatch2 T, outcome(_b) pscore(PSCORE)

The outcome is disapointing, I get _b that are all zero or missing.
I presume it could be that i have a lot of missing values amid the b  
variables, or maybe my specification is just wrong.
Do you have any experience on this?

Thanks!

Christian




*
*   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/


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