Re: st: DiD

[inggrid <inggrid_ep@yahoo.com>]

Hello Federico,

Thank you very much for your suggestion.
The results did not change much when I was omitting the weight and cluster.
You are correct, treat and treat_post05 are highly correlated.Should I drop "treat"? 

Indeed, I also created an interaction dummy between  treat_post05 and a covariate (X1). Hence, the full model is:

regress  y treat post05  treat_post05 treat_post05*X1 X2 X3 [aweight=hhweight],r cluster(ea)

Please correct me if I a wrong. I am just learning about the DiD. 

Kind regards,
Inggrid

--- On Thu, 3/31/11, federico.tedeschi@univr.it <federico.tedeschi@univr.it> wrote:

> From: federico.tedeschi@univr.it <federico.tedeschi@univr.it>
> Subject: Re: st: DiD
> To: statalist@hsphsun2.harvard.edu
> Date: Thursday, March 31, 2011, 3:51 PM
> So, you have this regression:
> 
> > regress  y treat post05  treat_post05
> [aweight=hhweight],r cluster(ea)
> 
> with a dataset like this:
> 
> > id year
> y   program   treat 
> post05  treat_post05
> 1 2000 10 0 0 0 0
> 2 2000 20 0 0 0 0
> 3 2000 30 0 0 0 0
> 4 2000 40 0 1 0 0
> 5 2000 50 0 1 0 0
> 6 2000 60 0 1 0 0
> 1 2007 70 0 0 1 0
> 2 2007 80 0 0 1 0
> 3 2007 90 0 0 1 0
> 4 2007 100 1 1 1 1
> 5 2007 110 1 1 1 1
> 6 2007 120 1 1 1 1
> 7 2007 130 0 0 1 0
> 8 2007 140 0 0 1 0
> 9 2007 150 0 0 1 0
> 
> and treat*post05 is dropped? 
> I think you could try to see how things changes if you
> change the weights, or omit treat (or post05) or the
> clusters, and also see how much treat*post05 is correlated
> with treat and post05.
> Federico  
> *
> *   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/