I'm using a difference-in-differences design to estimate the effect of an external shock on electoral turnout. (The treatment group received the shock, the control group did not, and the groups satisfy the parallel trend assumption.) I have eight total time periods (two elections prior to treatment, and six after treatment). I'd like to use this data to also estimate how long the external shock effects turnout in the treatment group. In other words, do we still register an effect after two elections? Three? Five? Is there a specific statistical technique I can use to test this? Thanks in advance.
Asked
Active
Viewed 14 times
2
-
You can probably find a significance test somewhere for this, but the first thing to do is to plot your data, look at them, and see if it seems plausible to make this sort of judgment in the first place. – user3716267 Jun 15 '21 at 15:56