Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: Re: Adding the marginal effects at individual values of


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Re: Adding the marginal effects at individual values of
Date   Mon, 22 Feb 2010 17:34:11 -0800 (PST)

--- On Feb 22, 2010, at 2:33 AM, Solomon wrote:
> This is important because
> I have a reason to believe that the marginal effects 
> will be different at high and low values of the regressor
> and the AME or the marginal effect at mean do not help me to
> verify this possibility.  

The problem with average marginal effects in this kind of 
application is that control variables no longer do what they
are supposed to do. The marginal effect will be large when the
predicted probability is close to .5, and will get less when the
probability moves closer to either 0 or 1. So, if the composition
of control variables changes over the variable of interest, then 
what you see in changes in average marginal effects consists both
of changes in the variable of interest _and_ of changes in the 
control variables. In one sense this captures real changes, in 
another sense it prevents us from comparing "like with like" as
suposedly you added the control variables for a reason, and they
are now no longer kept constant.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
-------------------------- 


      

*
*   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   |   Site index