In a difference-in-differences setting, if we deal with yearly data, we normally test the parallel trend for 3 or 4 years before the event dates. However, when dealing with daily data, how many dates that we need to test for parallel trends is what I am of curiosity about now. Is there any reference for it?
0 Response to How many pre-treatment periods we should test for paralell trend in the case of daily data?
Post a Comment