The time after did method is too short.

Economists often use counterfactual framework when evaluating the processing effect of a policy or event, and call the difference between the processed data (processing group) (actually observable) and the unprocessed data (counterfactual) "processing effect". Because the raw data is counterfactual and cannot be actually observed, the raw control group similar to the processing group in all aspects is often sought as a substitute for the counterfactual of the processing group, and the difference between the changes before and after the processing group and the changes before and after the control group is the "processing effect".

When estimating the processing effect, if the processing group has only one object to be processed, the integrated control method (SCM) is used. If the processing group has multiple processing objects, the difference among differences (DID or DD) is used; If the processing time is the same time point, the traditional DID is used. If the processing time is different, the multiphase DID is used.

DID is the most widely used measurement method to estimate the therapeutic effect. The idea of this method was first put forward by the medical scientist john snow (1855) when he was studying the cholera epidemic in London, and was introduced into economics when Aubeneau and von de Nienburg (19 15) were studying the influence of the minimum wage law. In order to estimate the therapeutic effect, compare the differences before and after treatment, and subtract the changes in the control group from the changes in the treatment group, which is the therapeutic effect, hence the name "double difference". The premise that DID needs to meet is that the treatment group and the control group must have the same trend to be treated, which is the hypothesis of "parallel trend" or "* * * same trend".

If the test finds that the hypothesis of * * * same trend is satisfied, the DID method can be directly adopted; If it is found that the assumption of * * * same trend is not satisfied, it is necessary to use the double difference method (PSM-DID) based on the trend score matching method. The basic idea of PSM-DID is to construct a control group with the same trend as the treatment group by using PSM method, that is, to select samples with the same or similar trend score as the actual control group in the total control group, so that the treatment group and the control group can meet the assumption of the same trend.