Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

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

Re: st: Testing equality across two specifications of ivreg2

From   Austin Nichols <>
Subject   Re: st: Testing equality across two specifications of ivreg2
Date   Wed, 5 May 2010 12:43:25 -0400

Jessica Ellen Leight <> :
This shows two specifications with the same sample:

On Wed, May 5, 2010 at 12:34 PM, Jessica Ellen Leight <> wrote:
> Dear all,
> I am estimating two different instrumental variables specifications using ivreg2
> on the same sample.  I have observations on two individuals per household (call
> them x1 and x2), each instrumented with an individual-level shock, and I
> estimate first the household fixed effects specification using the difference
> across individuals.
> ivreg2 ydif (xdif=shockdif), cl(village year)
> Subsequently, I estimate an individual's outcome as a function of both x1 and
> x2, appropriately instrumented, without household fixed effects.
> ivreg2 y1 (x1 x2=shock1 shock2), cl(county year)
> I now wish to test whether the coefficient on x2 in the second specification is
> equal to the coefficient on xdif in the first specification, or (a second test)
> whether _b[xdif]=_b[x2]-_b[x1].  Suest does not work smoothly with ivreg2 as
> others have noted, and previously suggested solutions on this list all seem to
> apply to cases in which the same specification is being estimated on two
> different subsamples, rather than two slightly different specifications with
> the same sample.
> Any suggestions are appreciated.
> Best,
> JL

*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index