面板數(shù)據(jù)做DID,基本的Regression Model為:
Y=a0+a1*Post+a2*treat+a3*Post*Treat+other controls;
其中Post和Treat為Dummy Variable
Post=1 for the pre-treatment period and =0 for the post-treatment period;
Treat=1 for the obs received treatment and treat=0 for obs not received treatment.
假設(shè)a,b,c,d接受了treatment,e,f,...z沒有接受treatment
其中a,b在1997年接受treatment, c,d 在1998年接受treatment,那么各變量的variable 應(yīng)該怎么放呢?
數(shù)據(jù)結(jié)構(gòu)如下
ID Year Post Treat
a 1997 0 1
b 1997 0 1
c 1997 0 0
d 1997 0 0
e 1997 0 0
f 1997 0 0
....
z 1997 0 0
a 1998 1 1
b 1998 1 1
c 1998 0 1
d 1998 0 1
e ????????
f ????????
... ????????
z ????????
我的問題是1998年的時(shí)候,e到z的post是1還是0呢,對(duì)于a和b來說,e到z是control,而且post是1,對(duì)于c和d來說,e到z也是control 但是post是0.
謝謝了!