The dependent variable is Hours Worked per week of youth during school
year. Since hours worked is censored at zero and 70 hours, I use the tobit
model.
What I'd like to do is to get predicted hours worked by race/ethnicity.
There are 4 groups (white/black/Hispanic/Asian), of which reference group
is white.
I'm wondering whether or not it's O.K. to make the same simulation as logit
or probit simulation. If there're other ways, please let me learn about
them.
