st: Re: Using the variables from one sw-result in another regression

["Michael Blasnik" <michael.blasnik@verizon.net>]

I think there might be an easier way, but this should work:

local cons "_cons"

local vars: colnames e(b)
local vars: list vars - cons

If you are looping in a do file, the first line only needs to occur once 
(before the loop).  The last two lines can be inside a loop and should pick 
off the names of the independent variables in the last estimation results 
and put them in the local macro `vars'

Michael Blasnik
michael.blasnik@verizon.net


----- Original Message ----- 
From: "Torbj�rn Lindquist" <torbjorn.lindquist@econ.umu.se>
To: <statalist@hsphsun2.harvard.edu>
Sent: Wednesday, October 19, 2005 3:59 PM
Subject: st: Using the variables from one sw-result in another regression



> I am trying to identify the variables to used for matching. To do this I am 
> using a method for identification of variables described in an article by 
> de Luna and Waerenbaum (2005). They suggest that one should make an OLS 
> regression of the outcome and use the significant variables in a probit on 
> the treatment indicator. The variables that are significant in the last 
> regression is affecting both treatment and outcome and they are therefor 
> relatively good variables to include in the matching.
>
> I use the sw command to see what variables passes the significance test. 
> The question is, is there some way I can get the list of variables from 
> the first regression into the second without having to type them in 
> manually? It would help alot since I will do this for a large number of 
> samples.
>
> Regards
> Torbj�rn Lindquist
> PhD Student
> Department of Economics
> University of Ume�
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/