Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

Re: st: Interpreting results from heckprob


From   "Garrido, Melissa" <[email protected]>
To   "'[email protected]'" <[email protected]>
Subject   Re: st: Interpreting results from heckprob
Date   Tue, 28 Jun 2011 14:24:34 +0000

Hi Florian,

It looks like what could be happening is that you have variables that are highly related to product innovations but not market novelties, and my guess is that product innovations and market novelties are highly correlated.  If you run a simple probit model, you're not accounting for these relationships so it appears that the explanatory variables are related to market novelties as well (the simple probit model masks some of the more nuanced relationships). 

To explore this further, you might consider calculating a variety of marginal effects after running -heckprob- (see Stata help for heckprob postestimation).  With these, you'll be able to calculate the marginal effect of your explanatory variables on selection (product innovations) and the marginal effect of your explanatory variables on your outcome conditional on selection, as well as other effects. 

Hope this helps,

Melissa


-------------------------------------------------------------------- 
Melissa Garrido, PhD 

Research Health Science Specialist 
GRECC/REAP, James J Peters VA Medical Center 
130 West Kingsbridge Road 
Bronx, NY 10468 
718-584-9000 x 3804 
[email protected] 

Assistant Professor 
Brookdale Department of Geriatrics & Palliative Medicine 
Box 1070 
Mount Sinai School of Medicine 
One Gustave L. Levy Place 
New York, NY 10029 
[email protected]



>st: Interpreting results from heckprob
>From 	  "Florian Seliger" <[email protected]>
>To 	  [email protected]
>Subject 	  st: Interpreting results from heckprob
>Date 	  Mon, 27 Jun 2011 14:59:30 +0200

>Dear Statalist,

>I'm trying to estimate a heckprobit model with CIS data.
>The dependent variable in the selection equation is whether a firm has introduced product innovations.
>The dependent variable in the outcome equation is whether a firm (that has introduced product innovations) has introduced market novelties.

>product innovation yes------->market novelty------>yes
                   no-------->-             ------>no

>I'm using a set of control variables such as the log of the number of employees, R&D activities, the proportion of employees with higher education etc., >the model variables I'm interested in, and different exclusion restrictions.

>The problem is that coefficients of the log number of employees, R&D activities, and the proportion of employees become insignificant in the outcome >equation in every specification I have tried out.
>Running a simple probit model, all coefficients are highly significant.

>My question is:
>- How can I interpret this result?
>- Do I use the wrong model?

>In addition, I would like to ask you if you have an idea which variable (from the CIS survey or additional data) to use as instrument in the selection >equation.

>Thank you!

>Best wishes,
>Florian
>-- 
>NEU: FreePhone - kostenlos mobil telefonieren!			
>Jetzt informieren: http://www.gmx.net/de/go/freephone



*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index