Quantcast

model checking after ivprobit

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

model checking after ivprobit

xueliansharon
Dear all,

I estimate a ivprobit model with two endogenous variables, then I want to check whether probit model is the appropriate model, i.e. whether the assumptions of probit model are satisfied. So do I need to check the normality of the residuals? How to do this test? And is there any other test I need to do?

Your help will be greatly appreciated.


Best,
Sharon
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: model checking after ivprobit

Conor Hughes
After you run the probit, why not just do something along the lines of
-predict r, resid-, and then plot it and compare visually using
-kdensity r, normal-?  I'm sure there are ways to get numerical
estimates of your fit, but this should give you a rough idea.

- Conor
On Sat, Sep 25, 2010 at 6:18 PM, xueliansharon <[hidden email]> wrote:

>> Dear all,
>
> I estimate a ivprobit model with two endogenous variables, then I want to
> check whether probit model is the appropriate model, i.e. whether the
> assumptions of probit model are satisfied. So do I need to check the
> normality of the residuals? How to do this test? And is there any other test
> I need to do?
>
> Your help will be greatly appreciated.
>
>
> Best,
> Sharon
> --
> View this message in context: http://statalist.1588530.n2.nabble.com/model-checking-after-ivprobit-tp5570987p5570987.html
> Sent from the Statalist mailing list archive at Nabble.com.
> *
> *   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/



--
Conor Hughes
Mathematics and Economics
University of Chicago 2011

*
*   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: model checking after ivprobit

xueliansharon
Conor,

Thanks for your suggestion. The problem is that I can't run "predict r, residuals" after probit, since option "residuals" is not allowed after probit estimation.

Any other idea?  And I would appreciate a lot if some numerical methods for testing normality of the residual could be provided.

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

Re: model checking after ivprobit

Conor Hughes
True, that was my mistake.  I overlooked that you were doing a probit,
which is integral to your question.  Anyway, you can get deviance
residuals in postestimation with probit in Stata, using the -deviance-
option with -predict-.  Deviance residuals would be much more suited
to probit regression than standard residuals anyway, in most cases.

- Conor

On Sat, Sep 25, 2010 at 9:31 PM, Conor Hughes <[hidden email]> wrote:

> True, that was my mistake.  But you can get deviance residuals
> postestimation with probit,  using the -deviance- option with -predict-,
> which would actually be more appropriate to use with probit.
>
> - Conor
>
> On Sep 25, 2010 8:52 PM, "xueliansharon" <[hidden email]> wrote:
>> Conor,
>>
>> Thanks for your suggestion. The problem is that I can't run "predict r,
>> residuals" after probit, since option "residuals" is not allowed after
>> probit estimation.
>>
>> Any other idea? And I would appreciate a lot if some numerical methods for
>> testing normality of the residual could be provided.
>>
>> Sharon
>> --
>> View this message in context:
>> http://statalist.1588530.n2.nabble.com/model-checking-after-ivprobit-tp5570987p5571170.html
>> Sent from the Statalist mailing list archive at Nabble.com.
>> *
>> * 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/
>



--
Conor Hughes
Mathematics and Economics
University of Chicago 2011

*
*   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: Re: model checking after ivprobit

Schaffer, Mark E
In reply to this post by xueliansharon
Sharon,

-overid- will report an overidentification test statistic after
-ivprobit-.

--Mark

NB: I suggest you include the preceding exchanges when you reply to
Statalist postings - it makes it much easier for the rest of us to work
out what you are referring to.

> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of
> xueliansharon
> Sent: 26 September 2010 02:53
> To: [hidden email]
> Subject: st: Re: model checking after ivprobit
>
> Conor,
>
> Thanks for your suggestion. The problem is that I can't run
> "predict r, residuals" after probit, since option "residuals"
> is not allowed after probit estimation.
>
> Any other idea?  And I would appreciate a lot if some
> numerical methods for testing normality of the residual could
> be provided.
>
> Sharon
> --
> View this message in context:
> http://statalist.1588530.n2.nabble.com/model-checking-after-iv
> probit-tp5570987p5571170.html
> Sent from the Statalist mailing list archive at Nabble.com.
> *
> *   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/
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: model checking after ivprobit

xueliansharon
In reply to this post by Conor Hughes
Conor,

Thanks for your recommendation. But the option "deviance" is not allowed after ivprobit,(it's allowed after probit), so do you have any other idea about computing the residuals and testing their normality?

Thanks,
Sharon



Quote: Sep 26, 2010; 01:54am — by Conor Hughes
True, that was my mistake.  I overlooked that you were doing a probit,
which is integral to your question.  Anyway, you can get deviance
residuals in postestimation with probit in Stata, using the -deviance-
option with -predict-.  Deviance residuals would be much more suited
to probit regression than standard residuals anyway, in most cases.

- Conor
Loading...