search

Daniel-Pailanir / sdid

Synthetic Difference in Differences for Stata
GNU General Public License v3.0
72 stars 36 forks source link

Graphs in Staggered Adoption #46

Closed mldoucette1 closed 1 year ago

mldoucette1 commented 1 year ago

Hello Daniel,

Thanks again for your quick response related to the SE in the e(tau) matrix, it was very helpful.

I have another question and appreciate your time. In the setting of staggered treatment adoption, is there a way to get 1 graph for both the trend line and/or the state weights? I tried combining the treatment cohorts using the e(difference) and e(series) matrices. Do they need to be weighted by treatment cohort length * size? Any thoughts would be very helpful.

Thanks,

Mitch Doucette mdoucet3@jhu.edu

damiancclarke commented 1 year ago

Hi Mitch,

If you want to do some non-standard graphical processing, we have laid out some examples about how to do this in a new working paper, here: https://docs.iza.org/dp15907.pdf. On pages 19-20, and appendix 2 we show how the graphs that are automatically exported can be made using the ereturned elements, so hopefully with this you can further edit graphs as you need!

Best wishes, Damian