Statalist


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

Re: st: Stata: IMR after biprob


From   rhythm Em <rhythmem@yahoo.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Stata: IMR after biprob
Date   Tue, 4 Aug 2009 14:11:09 -0700 (PDT)


Dear Austin Nichols,

After a couple of user written commands, the model you forwarded runs. Thanks for your response. 
I am sorry to be unclear about some of the things. Actually, I am searching for something like:
y1= 1 or 0
y2= 1 or 0 (y1 and y2 are partilaly related, partial observability of Meng and Schmidt type is appropriate) 

I formulated: biprobit y1 y2 x1 x2 x3 x4, partial

I want to have predicted probabilities and IMRs from first equation to put on my second equation like:

wage = y1 y2 x1 x2 x3 x4 x5 predicted-probs IMRs 

so that I can get appropriate cofficients for y1 and y2 in relation to wage in second equation. Note: wage is not binary. Now, does the command ivreg2 works in this case? can you suggest me the program to formulate in stata for this model. I am thankful for your all responses and hope to get help if you know. Thanks.

Sincerely,
Rhythm


----- Original Message ----
From: Austin Nichols <austinnichols@gmail.com>
To: statalist@hsphsun2.harvard.edu
Sent: Friday, July 31, 2009 7:11:46 PM
Subject: Re: st: Stata: IMR after biprob

rhythm Em<rhythmem@yahoo.com> :

The code runs if you cut and paste it into the command window, but
uses some user-written commands, so preface it with:

ssc inst ivreg2
ssc inst ranktest
ssc inst estout

but did you try

ssc desc cmp

?

On Fri, Jul 31, 2009 at 3:27 PM, rhythm Em<rhythmem@yahoo.com> wrote:
>
> Dear Austin Nichols,
> Thank you very much for your suggestion. I got some general idea to formulate the model even though I could not run the program
> you forwarded me (it says: lwage is not a valid estimator). Back to my model -- more preciously, I found that my dependendent variables are partially related. The option "partial" after biprob, I think is suitable for partial observability. Will model formulated in the way you suggested
> in previous e-mail also work in this case? I seek your suggestions regarding program formulation.Thank You.
> Regards,
> Rhythm
>
>
> rhythm Em<rhythmem@yahoo.com> :
> ssc desc cmp
> or try e.g.
>
> use http://fmwww.bc.edu/ec-p/data/wooldridge/card, clear
> g c=educ>15
> g m=married==1 if married<.
> ivreg2 lwage south smsa (c m=IQ black south66 reg66*)
> est sto iv1
> biprobit c m IQ black south66 reg66*
> predict p11
> predict p10, p10
> predict p01, p01
> g y1=p11+p10
> g y2=p11+p01
> ivreg2 lwage south smsa (c m=y1 y2)
> est sto iv181
> esttab iv1 iv181, nogaps mti eq(1) scal(widstat)
>
> On Fri, Jul 31, 2009 at 1:03 AM, rhythm Em<rhythmem@yahoo.com> wrote:
>> Hello Stata users,
>> I want to apply bivarite probit (as first stage) followed by OLS equation. In second stage, I want to include predicted probabilities and inverse mills ratio as regressor. I have formulated first stage using biprob as:
>> biprobit y1 y2 x1 x2 x3 x4, robust. Would someone help me to calculate inverse mills ratio after  biprobit model in stata. I went through the statalist archive about this type of problem but couldnot get the appropriate answer. So, can anyone suggest me to handle my situation?
>> I really appreciate your time and contribution.

*
*  For searches and help try:
*  http://www.stata.com/help.cgi?searchhttp://www.stata.com/support/statalist/faqhttp://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/



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