Topic stacked-did (on new flask repository)#1265
Topic stacked-did (on new flask repository)#1265valerievossen wants to merge 4 commits intomain-flaskfrom
Conversation
There was a problem hiding this comment.
Hi Valerie,
Great work! :D
I've added my suggestions.
As we discussed earlier, this alternative method to stacked DiD (Wing, C., Freedman, S. M., & Hollingsworth, A. (2024)) seems to be pretty new and is still a working paper. Would appreciate your take on this @kleintob: Should we keep this as is or should we maybe focus on the more established Stacked DiD method proposed by Cengiz et al. 2019? Feel free to add other suggestions you may have :D
|
|
||
| Other approaches to aggregating ATT values include: | ||
|
|
||
| - *Population-weighted ATT*: Weights the group-time ATT by its share of the treated population (instead of the treated sample) |
There was a problem hiding this comment.
Its a bit hard to give enough context on when/why to use the alternate weighting methods in a tip box alone. Maybe make it a subheading and give more info but m afraid the BB will get too long. To avoid info overload I'd simply point people to the relevant section in the paper for the different weighting approach. Its a whole discussion on its own ;)
There was a problem hiding this comment.
I get your point! Do you agree that leaving it in a tipbox and referring to the paper section (as I did now), is good for leaving the discussion but highlighting that there's alternative methods?
|
@srosh2000 Thank you for the valuable feedback!! I implemented your comments |
|
@srosh2000 can I merge this? Thanks!A |
(This is a new PR to the new master repository, the older one, #1236, I will close.
I checked the new style guidelines and fixed the links such that they work on the new website).
Hi @shrabasteebanerjee and @srosh2000,
This is the Stacked DiD topic, introducing the Weighted Stacked DiD method.
I kept the section about application in R small and highlighted the key steps, as the tutorial provided by the authors is really straightforward already and includes the function definitions.
Let me know what you think, thank you!