search

mcaceresb / stata-honestdid

Robust inference in difference-in-differences and event study designs (Stata version of the R package of the same name)
70 stars 18 forks source link

The problem of taking the value of M #15

Closed zkzhu-pro closed 1 year ago

zkzhu-pro commented 1 year ago

What should I consider about the value of M. Does M have to be equal to 1 or 2? I have applied this method. But when M_hat is greater than 1, the lower bound is less than zero, does this mean that DID results are not very good? the image is as follows: MM Thank you for your help！

jonathandroth commented 1 year ago

Hi, assuming that this is using the relative magnitudes restriction (the default in Stata), this suggests that the significance of the result is robust to allowing for violations of parallel trends as large as about 0.8 times the largest violation of parallel trends in the pre-treatment.

On Thu, Apr 13, 2023 at 11:20 AM zzk-19 @.***> wrote:

What should I consider about the value of M. Does M have to be equal to 1 or 2? I have applied this method. But when M_hat is greater than 1, the lower bound is less than zero, does this mean that DID results are not very good? the image is as follows: [image: MM] https://user-images.githubusercontent.com/59312438/231806295-2252a525-3b90-4d96-85e4-d0d3e5759bfa.jpg Thank you for your help！

— Reply to this email directly, view it on GitHub https://github.com/mcaceresb/stata-honestdid/issues/15, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE6EXFCKR5HEE2AFZK4GRW3XBAKT3ANCNFSM6AAAAAAW5HTIZQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>