Hi, I came across two commands that are supposed to do similar jobs: -kmatch- and -diff-
Both of them have propensity score matching. However, I can't find the way to make them compatible.
For example, from what I understand, if -time- is a dummy: 0=pre-intervention, and 1=post-intervention
and -treatment-: 0 = people not in the intervention, and 1=people in the intervention
then, the DiD (without matching) is basically:
reg outcome time#treatment
This is identical to:
diff outcome, p(time) t(treatment)
So far, everything worked.
The problem starts here. What if the assumptions of DiD are not satisfied, and I want to do a propensity matching? Then, the command -diff- can do it as follows:
diff outcome, p(time) t(treatment) ps( xvar1 xvar2 )
However, I can't get the same results using kmatch:
kmatch ps treatment xvar1 xvar2 (outcome)
Also, replacing -treatment- by TT= treatment*time still doesn't work.
Can someone please give me an easy solution with the two codes I need to make diff and kmatch compatible??
Thank you very much in advance,
0 Response to DiD and Matching
Post a Comment