Difficulties with the parallel trends assumption DiD

[Fergus McConkey]

Hi All,
I am currently attempting to model the effect of florida pill mill crackdown on opioid prescriptions using a DiD model with Georgia as my control state. I have county-monthly data for the years 2006 to 2019. I am using the Didregress command in stata and when i use the command estat trendplots my trends look reasonably parallel to the naked eye, as seen below, however, when i use the command estat ptrends i get absolutely tiny values for my p value, such as 0.0003, i was wondering if there was a reason for this or am i doing something wrong? Thanks in advance!

This is the trendplot using yearly data:

[Nick Cox]

Cross-posted at https://www.reddit.com/r/stata/comme...tion_violated/

Please note our longstanding request to tell us about cross-posting. This is a request here and a rule on Reddit's Stata subreddit.

[Fergus McConkey]

Sorry i didnt know about that rule, i will edit my post on reddit, my apologies.

[Jared Greathouse]

I don't follow. The trends, either way, don't look parallel to me. The trend is the treated group is clearly much steeper than the control group. Either way, what's your unit of analysis? Are you one comparing one state to another or is it more complicated?

[Carlo Lazzaro]

Fergus:
you might be luckier with synthetic control method, if you can collect a bunch of states to create a control.