Open vlukasin opened 2 years ago
I'm currently puzzling through this as well. My sense for the second part of the question is we do not want the coefficient to be statistically significant. If it is statistically significant then that means there is a difference, since we are talking about parallel lines we don't want a difference, we would want the slope to be the same.
It has to be a coefficient for a dummy variable since that is what allows us to compare the two groups.... But that's as far as I get....
The null hypothesis in the diff-in-diff is that T2 is the same as the counter factual.
if you think about it, the counter factual is the parallel line because it is just T1+trend.
So b3 is the test for program effects (we want a significant coefficient), but it is also the test for parallel lines (we do NOT want significance).
Hi, I am quite confused on which coefficient would capture the parallel lines test; can any one help pretty pleasseee??