Constructing DID model using pooled cross-section data

[Tiantian Liang]

Hi All:
I have a problem that how to use pooled cross-section data to construct a DID model.
Assume as follows, there are two schools A and B, each school has 10 different majors. Various information on graduates from 2014 to 2020 and the employment situation of each student were collected to form a pooled cross-section data set.
School A started a new policy in 2016: opening employment guidance classes. School B has never offered employment guidance classes. Students in school A are free to choose whether to participate in tutoring classes.
Therefore, the sample can be divided into 5 categories:

1. Students from school A before 2016 (no employment guidance classes);
2. Students from school B before 2016 (no employment guidance classes);
3. Students from school A participating in employment guidance classes after 2016;
4. 2016 Students from school A who did not participate in employment guidance classes after 2016;
5. Students from school B after 2016 (no employment guidance classes)

What I want to explore is what impact employment guidance classes have on students’ employment rates.
How should I construct the DID model?
My initial idea is to define school A as the treatment group and school B as the control group.But how should students in school A who do not participate in employment guidance classes be defined?

Hope someone can help me.

[Tiantian Liang]

I've resolved it.