Quantcast

st: Compare ivreg coefficients across models

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

st: Compare ivreg coefficients across models

Bart Vanneste
Dear all,

I have two instrumental variable regression estimated on two partially overlapping samples. I'd like to statistically compare a coefficient of model 1 with a coefficient of model 2. When I use suest to construct a common variance-covariance matrix, I get the following error message:
(ivregress): ivregress is not supported by suest
(ivreg2): unable to generate scores for model modelone suest requires that predict allow the score option r(322);


What should I do to compare two coefficients across iv models?

Earlier discussion of the same issue did not resolve the problem:
http://www.stata.com/statalist/archive/2009-07/msg01012.html
http://www.stata.com/statalist/archive/2005-04/msg00754.html

Any help is greatly appreciated,
Bart


     
*
*   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/
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: st: Compare ivreg coefficients across models

Austin Nichols
Bart Vanneste <[hidden email]>:
You can try stacking the models and clustering by observation which
should be equivalent to the -suest- approach, but you might also try
bootstrapping to confirm this gives the right answer (stratifying by
three groups: in sample 1 only, in sample 2 only, in both samples).

est clear
sysuse nlsw88, clear
ta south smsa
reg wage hours if south
est sto south
reg wage hours if smsa
est sto smsa
suest south smsa
est sto suest
preserve
expand 2
bys idcode: g n=_n
keep if (n==1&south)|(n==2&smsa)
g hours1=hours*!(n==1&south)
g hours2=hours*!(n==2&smsa)
reg wage hours? n, cl(idcode)
est sto stacked
restore
esttab south smsa suest stacked, nogaps mti
* confirm stacked SEs match -suest- SEs above and coefs match in all cases.
ivreg wage (hours=tenure union) if south
est sto south
ivreg wage  (hours=tenure union) if smsa
est sto smsa
preserve
expand 2
bys idcode: g n=_n
keep if (n==1&south)|(n==2&smsa)
g hours1=hours*!(n==1&south)
g hours2=hours*!(n==2&smsa)
g ten1=tenure*!(n==1&south)
g ten2=tenure*!(n==2&smsa)
g union1=union*!(n==1&south)
g union2=union*!(n==2&smsa)
ivreg wage (hours?=ten? union?) n, cl(idcode)
est sto stacked
restore
esttab south smsa stacked, nogaps mti


On Fri, Nov 27, 2009 at 3:09 PM, Bart Vanneste <[hidden email]> wrote:

> Dear all,
>
> I have two instrumental variable regression estimated on two partially overlapping samples. I'd like to statistically compare a coefficient of model 1 with a coefficient of model 2. When I use suest to construct a common variance-covariance matrix, I get the following error message:
> (ivregress): ivregress is not supported by suest
> (ivreg2): unable to generate scores for model modelone suest requires that predict allow the score option r(322);
>
>
> What should I do to compare two coefficients across iv models?
>
> Earlier discussion of the same issue did not resolve the problem:
> http://www.stata.com/statalist/archive/2009-07/msg01012.html
> http://www.stata.com/statalist/archive/2005-04/msg00754.html
>
> Any help is greatly appreciated,
> Bart

*
*   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/
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: st: Compare ivreg coefficients across models

Bart Vanneste
Thanks, Austin! This solves the problem.

Bart


----- Original Message ----
From: Austin Nichols <[hidden email]>
To: [hidden email]
Sent: Sun, November 29, 2009 4:56:52 AM
Subject: Re: st: Compare ivreg coefficients across models

Bart Vanneste <[hidden email]>:
You can try stacking the models and clustering by observation which
should be equivalent to the -suest- approach, but you might also try
bootstrapping to confirm this gives the right answer (stratifying by
three groups: in sample 1 only, in sample 2 only, in both samples).

est clear
sysuse nlsw88, clear
ta south smsa
reg wage hours if south
est sto south
reg wage hours if smsa
est sto smsa
suest south smsa
est sto suest
preserve
expand 2
bys idcode: g n=_n
keep if (n==1&south)|(n==2&smsa)
g hours1=hours*!(n==1&south)
g hours2=hours*!(n==2&smsa)
reg wage hours? n, cl(idcode)
est sto stacked
restore
esttab south smsa suest stacked, nogaps mti
* confirm stacked SEs match -suest- SEs above and coefs match in all cases.
ivreg wage (hours=tenure union) if south
est sto south
ivreg wage  (hours=tenure union) if smsa
est sto smsa
preserve
expand 2
bys idcode: g n=_n
keep if (n==1&south)|(n==2&smsa)
g hours1=hours*!(n==1&south)
g hours2=hours*!(n==2&smsa)
g ten1=tenure*!(n==1&south)
g ten2=tenure*!(n==2&smsa)
g union1=union*!(n==1&south)
g union2=union*!(n==2&smsa)
ivreg wage (hours?=ten? union?) n, cl(idcode)
est sto stacked
restore
esttab south smsa stacked, nogaps mti


On Fri, Nov 27, 2009 at 3:09 PM, Bart Vanneste <[hidden email]> wrote:

> Dear all,
>
> I have two instrumental variable regression estimated on two partially overlapping samples. I'd like to statistically compare a coefficient of model 1 with a coefficient of model 2. When I use suest to construct a common variance-covariance matrix, I get the following error message:
> (ivregress): ivregress is not supported by suest
> (ivreg2): unable to generate scores for model modelone suest requires that predict allow the score option r(322);
>
>
> What should I do to compare two coefficients across iv models?
>
> Earlier discussion of the same issue did not resolve the problem:
> http://www.stata.com/statalist/archive/2009-07/msg01012.html
> http://www.stata.com/statalist/archive/2005-04/msg00754.html
>
> Any help is greatly appreciated,
> Bart

*
*   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/
Loading...