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: create a comparison group
From
Austin Nichols <[email protected]>
To
[email protected]
Subject
Re: st: create a comparison group
Date
Tue, 1 Mar 2011 16:47:35 -0500
Jian AKA Vic Z <[email protected]>:
One way--compute a nonparametric propensity score e.g. by running a
logit of my230 (a dummy indicating your 230 firms) on narrowly defined
categories of industry, size, and listing age (and possibly
interactions of those categories) and then -predict double p-. It will
take some iteration to make sure that no categories give you perfect
prediction, and all predictions are far from 0 and 1. Once you've done
that, you can pick the 230 firms with the p closest to each of your
230, but what do you want to do if there are ties? One thing to do is
to randomly sort your data and just take the first match, but I would
think you would want to generate a few candidate datasets if ties are
common to assess the impact of randomness in matching. If you are
drawing without replacement, you will want to update a dummy
indicating ineligibility for matching as you go (presumably your orig
230 are also ineligible for matching). There are a variety of
programs for picking nearest neighbors for matching on SSC and other
archives, but it probably is best to proceed from first principles in
your case. You could also compute nearest matches in another metric;
see e.g. -mahapick- on SSC.
On Tue, Mar 1, 2011 at 3:58 PM, Vic Z <[email protected]> wrote:
> Dear Statalister:
> I have a problem with stata about how to create a comparison group for my study.
>
> Right now I have a sample with 230 firms. I simply need to create a comparison
> group with 230 other firms, which should be in the same industry, with similar
> size and similar listing age to my orginal 230 firms. After that I will group
> two sample together and run some regression. My problem is how I can tell stata
> to pull out those 230 firms from compustat to have a one-one matching with my
> orginal data sample so that I will have 230 pair firms. In each pair, the two
> firms will be almost identical.
> Thank you in advance.
> Jian
*
* 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/