# How many leads do we include for parallel trends testing in daily data (Difference-in-Differences)?

Regarding using generalized difference-in-differences, we normally use the joint null test for 2 or 3 years before the event year. For example, in this answer, @Thomas Bilach mentioned we may perform the joint null test for 3 years before the event year.

However, how do we handle this when we have a date variable and the data is at the date-level? How many dates before the event we should go to test for parallel trends?

• By saying you have a “date” you simply mean you observe outcomes at the day level, correct? Commented Sep 1, 2021 at 17:14
• @ThomasBilach, Yes, I am observing outcomes at the daily level Commented Sep 1, 2021 at 19:39

And lastly, a large $$p$$-value resulting from a joint null test doesn't "prove" parallel trends. In fact, you should interpret insignificance in this context as one piece of evidence supporting—not confirming—parallel trends. We can never really "confirm" anything, as this assumption is not directly testable.