Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | Luca Fumarco <luca.fumarco@lnu.se> |
| To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
| Subject | st: Thread-Index: AQHMtFo+KNDs77p3gESnnog4xmJOSQ== |
| Date | Tue, 6 Dec 2011 22:01:38 +0100 |
Dear Statalisters I have found the answer, and I share it with you (..maybe you still have some relevant comments on it): Let's say, as by previous literature, that the Var(e)=[exp(Z'*gamma)]^2 hetprob y B1x1 B2x2 B3x3, het(x1) [with het(.) I hypothesize the variance volatility is given by x1] mehetprob [I get the lnsigma2 of x1, that is the estimate of gamma1] then the mfx1=p(Y=1)*[ B^k - ZBwk]/[exp(gamma1*x1)] If x1 is a dummy var, then: -> the id for which x1=0 we have sigma=1 -> the id for which x1=1 we have sigma=exp(gamma1*x1) you can check your result with the postestimation command: predict sigmahat, sigma Luca Fumarco Ph.D. Student Linnæus University * * 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/