I am using
Code:
nehurdle
command in Stata12 (ssc install nehurdle). I want to implement a lognormal hurdle model as the log of my dependent variable seems to be normally distributed. I can't seem to find the option of lognormal with the command. Is it equivalent to the exponential option, or should I simply take a log of the dependent variable? How to handle the log of zeros (the dependent variable has many zeros) in that case?