Interpreting "pretrends" package

[Cat Santos]

Dear All,

I hope this message finds you very well.

I am writing because I am trying to evaluate if the parallel trends of my staggered DiD holds. My concern is that although my pretrend coefficients are substantially lower than my posttrend coefficients, I do have one significant pre-trend.

Therefore, I tried to use Roth's (2022) pretrends package to evaluate my pre-trends. However, I am having a very hard time interpreting the results of the output graph.

I specified my DV as binary (=0 for periods in which a country is a dictatorship; =1 for periods in which a country is a democracy); and also as a continuous value (democracy index). My treatment is binary and corresponds to the exposure to a democratization protest.

These are the resulting graphs for the binary DV:

And for the continuous DV:

I don't understand what would constitute a violation of parallel trends in these graphs. In general, I don't know how to interpret the position of the red and blue lines in relation to the CIs.

Thank you so much for your help.

Kind regards,
Cat

[Ashani Abayasekara]

Hi Cat,

I was just wondering if you figured out how to interpret these graphs, as I am having the same issue.

Thanks,
Ashani

[George Ford]

Zero is in all the confidence intervals prior to 0 time period, and don't see any meaningful trend, so you're good to go.