Hi,

I want to estimate a generalised probit model in which the dependent variable is wealth in quintiles. Among many dependent variables I would like to include the labour status of the reference person of a household which is coded as Rlab = 1(employee), 2(self-employed), 3(Retired) and 4(Other).

My question is whether it is correct to include Rlab as such or declare it in Stata as i.Rlab or I should recode it as a dummy for each category.

Thanks for your help.
Ilias