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

# re: Re: Re: st: Matching samples in Stata

 From "Ariel Linden, DrPH" To Subject re: Re: Re: st: Matching samples in Stata Date Thu, 4 Oct 2012 10:58:55 -0400

```Because the weight will account for the nature of the matches. If you have
1:1 matching, all the matched pairs will receive a weight of 1 (all
non-matched units will receive no weights and thus excluded from the outcome
model). With 1:1 matching you can forego the _weight and instead write the
regression as follows:

regress outcome treatment if _weight==1

If you use a different matching scheme (ie., kernel etc.), where every unit
in the data will receive a weight (or if you choose 1:k matching within a
caliper), then that weight must be included in the regression as I indicated
earlier:

regress outcome treatment [_weight]

Date: Wed, 3 Oct 2012 12:29:58 +0100 (BST)
From: Paula Arce <paulaarce@rocketmail.com>
Subject: Re: Re: st: Matching samples in Stata

Thanks for the clarification Ariel; why do I need to specify _weight in the
regression?

Cheers,
Paula

- ----- Original Message -----
From: "Ariel Linden, DrPH" <ariel.linden@gmail.com>
To: statalist@hsphsun2.harvard.edu
Cc:
Sent: Tuesday, 2 October 2012, 17:55
Subject: re: Re: st: Matching samples in Stata

The -outvar- in -psmatch2 is the outcome variable. In other words, if you
want to see if medical students have a significantly (statistically)
different outcome than non-medical students, you could plug that variable in
to -out()-, however the matching code will run perfectly fine without
specifying the outcome. That only means that you'll need to analyze the
outcome on your own after matching, ie:

regress outcome treatment [_weight]

Ariel

Date: Mon, 1 Oct 2012 15:39:59 +0100 (BST)
From: Paula Arce <paulaarce@rocketmail.com>
Subject: Re: st: Matching samples in Stata

Even after looking at the help command I am slightly unsure what I need to
specify where.

Should I first type the variable that codes the student type (1 for medical
0 for the rest) followed by the SES variables I want to match them  on?

psmatch2 student_type SES1 SES2, out(outvar)

Is the out(outvar) command going to produce a variable which lets me know
what participants from the 'other student' group are used as match for the
medics? if not how can I obtain this?

Thanks,
Paula

*
*   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/
```