Re: st: Re: Interaction and squared effects in a probit (pa) model

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Tue, 12/4/11, Laura Gomez-Mera wrote:
> Apart from the squared terms, we have other interaction
> effects of theoretical interest and we would like to be able
> to say something about the "conditional" marginal effects.
> Following Norton et al (2004) (http://www.unc.edu/~enorton/NortonWangAi.pdf)  we thought we could use Inteff, but since
> we are working with panel data, we are not sure this command
> is appropriate. 

A problem could be that panel models typically include a 
group-specific constant which is not directly estimated and the
marginal effects tend to depend on all parameters including 
these constants. There may be ways around that, but I am not
aware of any. I tend to prefer the logit link function and 
interpret odds ratios (the interaction terms are than ratios
of odds ratios). These will always be valid, as the odds ratio
does not depend on any of the other parameters in the model. I 
discussed this trick and an example of how interpret the 
results in:

Maarten L. Buis (2010) Stata tip 87: Interpretation of 
interactions in non-linear models. The Stata Journal, 10(2), 
305-308.  

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/