Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Re: IV and fixed effects


From   Jianhui Li <colasonic@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Re: IV and fixed effects
Date   Mon, 09 May 2005 10:36:10 -0300

Thank you guys for the help!

Mark: This is a survey data. every observation represents a certain number of
persons in the population. all the weights in the dataset addup to the whole
population. So the pweight is the weight I will use, isn't it?

Kit: Do I need to consider the weights while I am doing the demeaning? ie. when
I calculate the mean of a variable, should I use the sum of variables divided by
the number of oberservations, or should I use the sum of the products of the
variable and weight, then divide the sum by the total weights? I hope I've made
myself clear.


-----------------------------------------------------
Jianhui Li <colasonic@gmail.com>


Quoting mail from "Mark Schaffer" <M.E.Schaffer@hw.ac.uk> at Mon, 09 May 2005 12:26:14 +0100

Subject: Re: st: Re: IV and fixed effects

> Just to follow up on Kit's email,
> 
> From:           	Kit Baum <baum@bc.edu>
> Subject:        	st: Re: IV and fixed effects
> Date sent:      	Mon, 9 May 2005 07:11:47 -0400
> To:             	statalist@hsphsun2.harvard.edu
> Send reply to:  	statalist@hsphsun2.harvard.edu
> 
> > Presumably you have too many fixed effects to implement directly. But 
> > you can always apply the demeaning transformation "by hand" to each 
> > variable--that is, by unit, express each variable as a deviation for 
> > the unit's mean--and then apply any appropriate estimator to the 
> > transformed variables. That is all that the xtivreg,fe estimator (or 
> > any estimator of the LSDV model, such as xtreg, fe or areg) is doing 
> > anyway. The egen mean function (which supports the bys unit: prefix) 
> > may be useful in doing this demeaning transformation.
> 
> Ben Jann's -center- will do this very nicely, with a casewise option 
> and support for aweights.  Dealing with weights might be a bit 
> tricky, though - is it simply a case of demeaning using weighted 
> means?  If so, aweights is good enough.
> 
> --Mark
> 
> > 
> > Kit Baum, Boston College Economics
> > http://ideas.repec.org/e/pba1.html
> > 
> > On May 9, 2005, at 2:33 AM, Juanhui wrote:
> > 
> > > My survey data requires me to run fixed effect OLS regression with 
> > > instrumental
> > > variables. However, xtivreg doesn't support weights. ivreg doesn't 
> > > support fixed
> > > effects and areg doesn't support iv. Is there any way to combine them 
> > > together
> > > and get me finish my thesis?
> > 
> > *
> > *   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/
> 
> Prof. Mark E. Schaffer
> Director
> Centre for Economic Reform and Transformation
> Department of Economics
> School of Management & Languages
> Heriot-Watt University, Edinburgh EH14 4AS  UK
> 44-131-451-3494 direct
> 44-131-451-3296 fax
> 44-131-451-3485 CERT administrator
> http://www.sml.hw.ac.uk/cert
> 
> *
> *   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/

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index