Reporting results from xtdidregress

[Thibaud Marcesse]

Hello,

I am running a DiD with panel data and have been using the xtdidregress command (one treatment administered once). I am debating which results to report in a paper. I am able to use postestimation commands to generate Trend plots and Granger plots, but the results are different from the actual coefficients of the DiD. The Granger plot includes leads and lags. In my case, however, the difference between raw coefficients and time specific estimates in the Granger plot are quite large so I am wondering which results I should really emphasize (as the difference between the two would change the conclusions from my paper).

Thanks for letting me know!

[Carlo Lazzaro]

Thibaud:
see explanations of equations (3) and (4) -Methods and formulas-, -didregress postestimation- entry, Stata .pdf manual.

[Thibaud Marcesse]

Thank you Carlo!