Hi all,
We are doing a difference-in-difference model based on PSM and NN-approach. The treatment takes place between 2015 and 2018, and we ended up with a balanced panel with 796 matches. We have 749 control firms, meaning that some control firms have been matched multiple times (matching with replacement). Thus, some of the control firms have been matched with treatment firms multiple times in a year. Example: Firm A was matched with treatment firm 1 and 2 in 2015. So, when running the diff-in-diff model, we should account for this by using firm A as control firm twice in the DID-test, right?
Best,
Oliver and Philip
We are doing a difference-in-difference model based on PSM and NN-approach. The treatment takes place between 2015 and 2018, and we ended up with a balanced panel with 796 matches. We have 749 control firms, meaning that some control firms have been matched multiple times (matching with replacement). Thus, some of the control firms have been matched with treatment firms multiple times in a year. Example: Firm A was matched with treatment firm 1 and 2 in 2015. So, when running the diff-in-diff model, we should account for this by using firm A as control firm twice in the DID-test, right?
Best,
Oliver and Philip
