Statalist


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

Re: AW: st: RE: How to understand the linear prediction after -heckman-?


From   gjhxmu@sina.com
To   statalist<statalist@hsphsun2.harvard.edu>
Subject   Re: AW: st: RE: How to understand the linear prediction after -heckman-?
Date   Wed, 30 Sep 2009 00:19:14 +0800

Dear Martin,
thank you for your help.
I am sorry for my unclear expression.

I want to get a R-squared after -heckman- comparable to the OLS regression.
In the thread http://n2.nabble.com/st-Re-Stata-Heckman-and-Mills-ratio-tt3707672.html#a3707672, Kit refered to  "you can always create an r^2 measure from the squared correlation of   
actual and predicted observations used in the estimation",which I have known from Wooldridge's book before.

Following that, I thougt no matter calculate R-squared from prediction after -heckman- or 
from prediction after -reg- with mill the results were the same.

However,Previously I found they were different and after your help I knew it was something to do with the "mill". I know the result of -heckman- is equal to the result of calculation manually in which first do -probit- then -reg- with mill. Why their predictions are different? Namely, why not include "mill" in the prediction?

Only understanding the doubt about the prediction can I manually calculate R-squared.


Thank you very much.


Best regards,
Rose.

----- Original Message -----
From: Martin Weiss <martin.weiss1@gmx.de>
To: <statalist@hsphsun2.harvard.edu>
Subject: AW: st: RE: How to understand the linear prediction after -heckman-?
Date: 2009-9-29 19:14:42


<> 

I am not quite sure what the question really is. If you want some
pseudo-r-squared, look at Nick`s
http://www.stata.com/support/faqs/stat/rsquared.html

-predict- after -heckman- is quite comprehensive. What do you want in
addition to what is already available?

Also note the Cameron and Trivedi (2009) book, which deals with -heckman-
quite comprehensively. http://www.stata-press.com/books/mus.html
See chapter 16.6


HTH
Martin


-----Urspr黱gliche Nachricht-----
Von: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von gjhxmu@sina.com
Gesendet: Dienstag, 29. September 2009 11:22
An: statalist
Betreff: Re: st: RE: How to understand the linear prediction after
-heckman-?

Dear Martin,

thank you for your reply and help.

Indeed, after leaving "mill" out of my -regress- prediction, the results are
the same.

However, which one is correct? To add "mill" in the -heckman- prediction or
to leave it out of my -regress- prediction if I need to get the prediction
and R-squared after -heckman-?

Best regards,
Rose


----- Original Message -----
From: Martin Weiss <martin.weiss1@gmx.de>
To: <statalist@hsphsun2.harvard.edu>
Subject: st: RE: How to understand the linear prediction after -heckman-?
Date: 2009-9-29 03:03:45


<>

You have to leave "mill" out of your -predict-ion...


**************
use http://www.stata-press.com/data/r11/womenwk.dta, clear

heckman wage educ age, /*
*/ select(married children educ age)/*
*/ twostep mills(mill)
predict yheck

reg wage educ age mill
gen yreg=_b[_cons]+_b[educ]*educ + /*
*/ _b[age]*age

su y*
compare yheck yreg
**************


HTH
Martin


-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of gjhxmu@sina.com
Sent: Montag, 28. September 2009 19:38
To: statalist
Subject: st: How to understand the linear prediction after -heckman-?

Dear statalists,
In my mind, the result of -heckman- is the same as the result of -reg- with
lambda as an additional independent variable. However, I found something
different.
I typed the following in the stata, and the related results are as
followings.


. qui heckman wage educ age, select(married children educ age) twostep
mills(mill)

. qui predict double yheck if wage!=.

. qui corr yheck wage if wage!=.

. di r(rho)^2
.2512235

. qui reg wage educ age mill if wage!=.

. qui predict double yreg if wage!=.

. qui corr yreg wage if wage!=.

. di r(rho)^2
.27934983

. su yheck yreg

Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
yheck | 1343 21.96174 3.717985 14.79669 32.88486
yreg | 1343 23.69217 3.332614 16.22861 33.78897


The predicted value should be the same and the R-squared should be the same
too.


Why different? 

Hope for any help and thank you very much.



Best regards,

Rose.


*
* 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/

*
* 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/

*
*   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