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]

st: Difference in difference and non-linear regression model


From   Thomas Werner <thomas.werner@unibw.de>
To   statalist@hsphsun2.harvard.edu
Subject   st: Difference in difference and non-linear regression model
Date   Wed, 14 Nov 2012 17:26:18 +0100

Dear Statalisters

I want to see if a statutory minimum wage had any effect on the probability of price increases respectively decreases after its implementation.
The dataset I use contains among others monthly firm-level price (Y) 
state of business (S) and demand data (D) of business surveys.
Most of the variables have three categories, whereby the answers are 
coded as 1 ("increased"), 0 ("not changed") and -1 ("decreased").
The independent variables are recoded as binary variables but there are 
also continuous control variables like capacity utilization.
I have to run a stereotyped ordered logit model because the parallel 
regression assumption of the ordered logit model is violated.
Actually I want to run a diff-in-diff model like this:

slogit = treat + post + treat*post + lag_price + D_increase + D_decrease + S_increase + S_decrease + material_price + ... + error
But I don’t know how to compute the marginal effect of the interaction 
term (treat*post) because the inteff command doesn’t work as a result of
the trichotomic shape of the dependent variable and the margins command 
isn’t a viable alternative.
As far as I can see the interaction effect should be equal to the 
difference of cross differences
Is there a way to compute these effect with stata?

Best
Thomas Werner

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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