Double Sample Selection Heckman Model

[Duy To]

Hi everyone,

My data includes information for part-time, full-time workers and unemployed, and I am looking to examine the incidence of over-education among full-time workers. To address the sample selection bias, I consider using the Heckman Sample Selection technique, in which individuals are selected to be full-time.

However, only employed workers have the probability of working for a full-time job and being over-educated. Therefore, I think it is reasonable to consider another selection, which is a selection into employment. In short, firstly, individuals are selected into employment, then employed workers are selected into full-time employment (i.e., two selections in total).

With a single selection, I could use the -heckman- command. However, I do not know where to start with the two selections I mentioned above. Therefore, I would appreciate it if you guys could help me with this.

Thank you for your help in advance.

[Xiao Jun]

i have the same question

[FernandoRios]

Short answer: look into CMP. It allows you to estimate heckman type of selection models, including double selection.
Long answer: You may need to write from scratch the implications of a double selection, including the implications of the joint distributions, and partial distributions, so you can implement it using -ml-