Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Alexandros Zangelidis <a.zangelidis@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Mfx after zero-inflated (ZOIB) estimator |
Date | Mon, 4 Jun 2012 13:57:42 +0100 |
Dear STATAList users, I’m writing regarding the zero-one inflated beta distribution (ZOIB) estimator (as available from SSC). I’m currently using this estimator in a model of sickness absence in order to separate the process that determines (a) the decision of being absent (binary decision), from (b) the level of sickness absence from work (proportion of contracted hrs). After the estimation of the model using ZOIB and estimate the marginal effects: • mfx, predict(pr0) • mfx, predict(prcond) I was wondering whether you could help me understand how exactly the marginal effects are calculated (formula) and how comparisons can be made to the marginal effects in a standard probit/logit. For example are the marginal effects for each binary explanatory variable calculated by fitting two separate probability functions? Exactly how are the marginal effects for the continuous variables calculated? Looking forward to hearing from you, Kindest regards, Alexandros Dr. A. Zangelidis | Lecturer in Economics | University of Aberdeen | Scotland | UK * * 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/