Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: st: RE: psmatch2 and p-value
From
"Colette.Grey" <[email protected]>
To
<[email protected]>
Subject
RE: st: RE: psmatch2 and p-value
Date
Fri, 3 Dec 2010 15:44:51 -0000
DVM,
Thanks for your response; what I have done is include all the variables
- the selection of which was based on theory, institutional knowledge
and prior research. I just wondered after seeing the post if I was
correct.
Regards, Colette
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Dimitriy V.
Masterov
Sent: 02 December 2010 14:47
To: [email protected]
Subject: Re: st: RE: psmatch2 and p-value
Colette,
My understanding is that there is no deterministic variable selection
procedure that will tell you which variables to choose from those
available to do matching. However, more is better is not quite right.
You should try to include all variables that affect both participation
and the outcome variable. This information comes from theory,
institutional knowledge, and previous research. Including instruments
- variables that affect participation and not outcomes - is not a good
idea. They do not help with selection bias and may worsen the support
problem. I am not sure if this is a consensus view. Perhaps others can
chime in.
DVM
*
* 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/
*
* 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/