Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | "Dimitriy V. Masterov" <dvmaster@gmail.com> |
| To | Statalist <statalist@hsphsun2.harvard.edu> |
| Subject | st: two questions about tpm and percentage impacts |
| Date | Wed, 17 Jul 2013 16:37:34 -0700 |
1) Is this the correct way to estimate the percentage impacts using the Belotti-Deb command tpm? Here's a toy example: webuse womenwk, clear replace wage = 0 if wage==. tpm wage educ age i.married children, f(probit) s(glm, fam(gamma) link(log)) margins, dydx(married) margins, eydx(married) The first AME is $2.72. The second semi-elasticity is .2291251, so if I do the usual transformation, I get exp(.2291251)-1=0.25749934, which corresponds to am ~26% effect. 2) I used the gamma family and the log link above (as suggested in the SJ paper as an alternative to OLS with logged outcome). This seems strange to me in light of the use Poisson, tell a friend recommendations to use poisson (glm with a log link and Poisson family). DVM * * 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/