Difference-in-difference with different pre - and post - periods

[fama]

Hi every one,

I have a question regarding DID estimation model.

I created a data set using propensity score matching (psm) for observations that a same treatment happened to them in different period of time. If the treatment happened in period 2 we compare outcome of period 3 and 1 in treated and matched control groups.

recall treatment only happened in period2, 3 and 4.

Now i want to use DID model using proc glm or proc mixed etc to investigate the effect of treatment on outcome.

your help will be appreciated.

Thanks

OBS   perid1   period2   period3  period 4   period 5   treatment  MatchID  outcomes for all periods and other covariates

 1         1             2             2         .                 .                  yes

2           1             1            1                                                NO

3

.

.

[fama]

Hi every one,

I have a question regarding DID estimation model.

I created a data set using propensity score matching (psm) for observations that a same treatment happened to them in different period of time. If the treatment happened in period 2 we compare outcome of period 3 and 1 in treated and matched control groups.

recall treatment only happened in period2, 3 and 4.

Now i want to use DID model using proc glm or proc mixed etc to investigate the effect of treatment on outcome.

your help will be appreciated.

Thanks

OBS   perid1   period2   period3  period 4   period 5   treatment  MatchID  outcomes for all periods and other covariates

 1         1             2             2         .                 .                  yes

2           1             1            1                                                N

[Kurt_Bremser]

I merged your posts and moved them to the Statistical Procedures community.

[fama]

Thanks