Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: reg "partial" option equivalent to ivreg2

From   "Schaffer, Mark E" <[email protected]>
To   <[email protected]>
Subject   st: RE: reg "partial" option equivalent to ivreg2
Date   Wed, 27 Oct 2010 17:46:54 +0100


> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of 
> john sanders
> Sent: 27 October 2010 16:08
> To: [email protected]
> Subject: st: reg "partial" option equivalent to ivreg2
> exogenous variables before estimating the full model. This is 
> especially useful when you have many non-nested fixed 
> effects. If you are not interested in the coefficients of 
> these fixed effects per se, then there is not much sense in 
> estimating them. Think of a model with worker, firm, and 
> industry*year FE for which areg could knock out workers (or 
> perhaps worker*firm), but industry*year still creates a lot 
> of extra parameters to be estimated.
> The "partial" option is simply an application of the Frisch 
> Waugh Lowell theorem, but it helpfully uses a "degrees of freedom"
> correction to make the standard errors equivalent to 
> estimating the model in a single step.
> Is there a user written command for a linear regression that 
> has this partial option?

Yes - it's called -ivreg2-!  -ivreg2- does simple linear regression as
well as IV regression.  The syntax is the same as that for -regress-.


> I guess I could pull the code out of 
> ivreg2.ado, but I thought I would ask here first.
> -John Sanders
> *
> *   For searches and help try:
> *
> *
> *

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

*   For searches and help try:

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