Re: st: Difference-in-differences analysis with binary data (repeated cross-sectional data)

[Armen Martirosyan <armen_martirosyan@wvi.org>]

Dear Maarten,

Thank you very much for your support. Regarding interpretation of results 
- how results should be interpreted if i get coefficient as a result of 
linear regression since for outcome binary data i am interested in 
proportions. 

Thank you for your consideration. Armen
 
___________________________________
Armen Martirosyan
Health Research and M&E Associate 
Global Health and WASH
World Vision 
International                                                                              

tel:   +374 10 749 118, 749 119
cell: + 374 96  36 81 89 
fax:  +374 10  749 148
skype: armen_mol
armen_martirosyan@wvi.org 
 
 



From:   Maarten Buis <maartenlbuis@gmail.com>
To:     statalist@hsphsun2.harvard.edu, 
Date:   05/24/2012 03:49 PM
Subject:        Re: st: Difference-in-differences analysis with binary 
data (repeated cross-sectional data)
Sent by:        owner-statalist@hsphsun2.harvard.edu



On Thu, May 24, 2012 at 1:19 PM, Armen Martirosyan wrote:
> I am using Stata 12 and doing Difference-in-differences (DiD) analysis
> with repeated cross-sectional data. I have found Stata syntax for DiD
> regression for continuous data but I am not able to find syntax for 
binary
> data and most of my data are binary. I would appreciate if you can send 
me
> Stata syntax for binary DiD analysis (repeated cross-sectional) along 
with
> regression equation and  some background information and examples (if
> available)  which can help me to understand how to deal with binary DiD
> regression.

If you have no additional control variables than you have a fully
saturated model and there is no problem with using the code for
continuous data with the addition of the -vce(robust)- option. See:
<http://www.stata.com/statalist/archive/2012-02/msg00351.html>

If you have additional control variables you need to be a bit more
careful and do some more checking to see if the model is reasonable,
but in many cases I would suspect that the linear probability model
will still be just fine.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------

*
*   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/