Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Re: RE: compute elasticity: Probit VS LPM


From   "Maarten Buis" <M.Buis@fsw.vu.nl>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Re: RE: compute elasticity: Probit VS LPM
Date   Tue, 18 Apr 2006 18:16:41 +0200

Feng:
What is causing the difference between linear approximations (OLS and your method) and probit is the very strong effect of x1. One way to see this is to look at the predicted probabilities for different values of x2 and x1, which you can get by using -prgen- from -spost- and create a graph of the predicted probability versus x2. You can also plot your linear approximation using the -twoway function- command and compare the two. See example code below. You can see that you have a very extreme example and that you should not trust results from OLS in this case.
HTH,
Maarten

*-----------begin example-------------
cd h:\temp
use teste.dta, clear
probit y x1 x2
prgen x2, x(x1=1) gen(x21)
prgen x2, x(x1=0) gen(x20)
reg y x1 x2
twoway line x21p1 x20p1 x21x || /*
  */ function y = _b[_cons] + _b[x1] + _b[x2]*x, range(1.3 3.7) || /*
  */ function y = _b[_cons] + _b[x2]*x, range(1.3 3.7) /*
  */ legend(label(1 "probit" "x1=1") label(2 "probit" "x1=0") /*
  */        label(3 "OLS" "x1=1")    label(4 "OLS" "x1=0") )
*----------------------end example--------------------------


-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology 
Vrije Universiteit Amsterdam 
Boelelaan 1081 
1081 HV Amsterdam 
The Netherlands

visiting adress:
Buitenveldertselaan 3 (Metropolitan), room Z214 

+31 20 5986715

http://home.fsw.vu.nl/m.buis/
-----------------------------------------

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu]On Behalf Of Feng Liu
Sent: dinsdag 18 april 2006 17:44
To: statalist@hsphsun2.harvard.edu
Subject: st: Re: RE: compute elasticity: Probit VS LPM

Usually I use Probit or Logit instaed of LPM. This time I guess the
elasticity from mfx,eyex after Probit might be wrong, so I tried LPM. Is
there a way to verify which one is right?
What I tried is:
after Probit regression,
.predict p0
.replace x2=x2*1.1
.predict p1
Then I compare p0 and p1 and find it fell by 3%. This implies the elasticity
is -0.3, which is close to the one from LPM.

 

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   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