Bookmark and Share

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]

st: RE: RE: testing coefficients across different ivreg models - large dataset with many variables


From   "Imberman, Scott A" <[email protected]>
To   "[email protected]" <[email protected]>
Subject   st: RE: RE: testing coefficients across different ivreg models - large dataset with many variables
Date   Thu, 10 Mar 2011 11:37:07 -0600

Thank's Mark.  It didn't occur to me to de-mean the data first.  That should work much faster.

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Schaffer, Mark E
Sent: Thursday, March 10, 2011 11:27 AM
To: [email protected]
Subject: st: RE: testing coefficients across different ivreg models - large dataset with many variables

Scott,

You can stack the data as per Austin's posting to Statalist, and use the partial option of -ivreg2- to partial out the RHS fixed effects.

Or - probably faster - you can use the within transformation (Ben Jann's -center- command is good for that) to demean your data and then stack.

HTH,
Mark

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of 
> Imberman, Scott A
> Sent: 10 March 2011 15:23
> To: '[email protected]'
> Subject: st: testing coefficients across different ivreg 
> models - large dataset with many variables
> 
> Dear Statalisters,
> 
> I'm trying to run the following models on two different subsamples
> 
> ivreg y = (x = z), cluster(k) if j = 1  (model1) ivreg y = (x 
> = z), cluster(k) if j = 2  (model2) ivreg y = (x = z), 
> cluster(k) if j = 3  (model3) ivreg y = (x = z), cluster(k) 
> if j = 4  (model4)
> 
> and I want to test if x(model 1)  = x(model 2) = x(model 3) = 
> x(model 4)
> 
> I know that I can't use suest with ivreg and I can't use 
> hausman due to the clustering.  I'm trying to stack the data 
> and run a fully interacted model as suggested in this posting 
> 
> http://www.stata.com/statalist/archive/2010-04/msg00366.html
> 
> but my dataset is very large and I have many fixed effects 
> and controls - about 1300 in each model, so 5200 in the fully 
> interacted model and so it's taking a very long time to run.  
> Is there any other way to test the coefficients across the 
> separate models?
> 
> Thanks.
> 
> Scott A. Imberman
> Assistant Professor of Economics
> University of Houston
> 204 McElhinney Hall
> Houston, TX 77204-5019
>  
> 713-743-3839
> [email protected]
> http://class.uh.edu/faculty/simberman
> 
> *
> *   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/
> 


-- 
Heriot-Watt University is a Scottish charity
registered under charity number SC000278.


*
*   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index