I have constructed an ordered logit model consisting of a latent variable with 3 responses and 4 explanatory variables as follows:
Y*= a + B1X1 + B2X3 + B3X3 + B4X4
My questions are:
1- What is the command used in EViews8 to calculate the average marginal effects of the explanatory variables?
2- Are these marginal effects single value or one value for each response level?
3- What is also the command for the marginal effects for the whole sample?
Your accurate answer is highly appreciated
Hello Abid, you can see this website: http://forums.eviews.com/viewtopic.php?f=4&t=337
Consider using STATA software for such models. It is easy to use and command driven. After running your basic logit model, you then enter a command "mfx" and the marginal effects will be computed. With Eviews, have never tested it, but have also run such models. You can check this paper; http://ssrn.com/abstract=2412639
http://ssrn.com/abstract=2412639
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics