Hello,
I wanted to ask the question in the other thread ([1]) but it does not seem to attract attention. I also think it could be of interest for a new topic. Please let me know if it is not appropriate.
I am trying to compare two non-nested models: Cragg's Tobit after -craggit- and Heckman sample selection model after -heckit-. I am aware of Santos Silva, Tenreyro, and Windmeijer (2015) after command --hpc-- but I would like to know how to do the Vuong's (1989) test for non-nested models. I understood the test for strictly nested models (equivalent to a LR test after command --lrtest--) and Cameron and Trivedi (2005) provide some codes for overlapping models in [2], however I am struggling to construct the Vuong's test for a Cragg's and a Heckit model.
Any instructions or sources are very much appreciated. I have read the reference paper Vuong (1989, Econometrica) but struggle to apply.
Thank you.
[1] http://www.statalist.org/forums/foru...nnested-models
[2] http://cameron.econ.ucdavis.edu/mmab...2nonnested.txt
I wanted to ask the question in the other thread ([1]) but it does not seem to attract attention. I also think it could be of interest for a new topic. Please let me know if it is not appropriate.
I am trying to compare two non-nested models: Cragg's Tobit after -craggit- and Heckman sample selection model after -heckit-. I am aware of Santos Silva, Tenreyro, and Windmeijer (2015) after command --hpc-- but I would like to know how to do the Vuong's (1989) test for non-nested models. I understood the test for strictly nested models (equivalent to a LR test after command --lrtest--) and Cameron and Trivedi (2005) provide some codes for overlapping models in [2], however I am struggling to construct the Vuong's test for a Cragg's and a Heckit model.
Any instructions or sources are very much appreciated. I have read the reference paper Vuong (1989, Econometrica) but struggle to apply.
Thank you.
[1] http://www.statalist.org/forums/foru...nnested-models
[2] http://cameron.econ.ucdavis.edu/mmab...2nonnested.txt