re: st: Which matching command works best?

["Ariel Linden, DrPH" <ariel.linden@gmail.com>]

Hi Benjamin

This seems to be a problem that can be solved with coarsened exact matching
(cem), a user written program (-findit cem-) that will allow you to set the
range of values for the matching variable(s).  This should address your
desire for matches " ...if they do not share identical values but also
if they are very likely to be similar..."

I hope this helps

Ariel

Date: Sat, 11 Feb 2012 13:25:27 +0100
From: Benjamin Niug <benjamin.niug@googlemail.com>
Subject: st: Which matching command works best?

Hello folks,

I have the following matching problem: I want to match observations in
2008 with observations in 2010 based on the following characteristics:
	
1.	Below poverty rate
2.	formal vs. self-employed
3.	male or female

I.e. the structure looks as follows:

id	year	   Poverty          formal_employed		male
1	2008	   1			1
1
2	2008	   0			0
0
1	2010	   1			1
1
2	2010	   0			0
1

As I have only about 50 observations per year, I would like to match
the observations only if they do not share identical values but also
if they are very likely to be similar (ie. the fourth and the second
observation should be merged although they differ only w.r.t. the male
variable).

I came accross the ccmatch .ado which might be applied in my case.
However, I was wondering whethere there is a more approprite matching
procedure for my specific problem out there?

Thanks in advance.
Benjamin
*
*   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/