On this note, if you ever want to use staggered SDID, check the appendix of this paper as we provide step-by-step derivation + intuition.

The #DiD #Winter2023 update is here with tons of new packages and code. It is still being updated but a LOT of new info is now there.

New packages: jwdid, wooldid, lpdid, fect, xtevent, sdid, hdidregress + updated to existing ones.

asjadnaqvi.github.io/DiD/

i'm pretty sure this (when you strip out the talk about the different models) says that the "iPhone counties" started declining faster (to catch up with "non-iPhone counties") before the iphone came out.

Thrilled to share that my first academic paper is out!

We investigated the impact of privatization on sanitation outcomes in Brazilian municipalities.

Grateful for my co-authors Ana Carolina Zoghbi and José Guilherme!

www.tandfonline.com/eprint/YI7I9...

Feeling uncertainty about the state of creative work? Lots of good nuggets in this conversation on approaching technology with intention from Adobe. https://adobe.design/stories/community/adobe-design-talks-navigating-uncertainty-in-creative-work?sdid=49F59DX9&mv=social

WHOLE ONION. Okay to be fair, when my parents were in the US they sdid say some onions are sweeter than others. I think maybe depends on the type. Onions here, I'd die chewing on it like an apple.

New @iza.org Working Paper, with @johnspringford.bsky.social docs.iza.org/dp18478.pdf "The Impact of Brexit on UK Immigration and Labour Supply: Evidence from Synthetic Differences in Differences"

For example, applying SDID separately to Spanish parties in the context described by Bolet et al. (APSR 2024), we get estimates similar to the DID ones in size, but their sum is not zero. With common weights, the sum constraint holds, while both the PSOE gain and the VOX loss attenuate.

a explicação é a diminuição da interação na vida real e maior acesso à pornografia, principalmente. queria entender melhor como lidaram com os confundidores. usar SDID aparentemente evita q o efeito seja atribuído de forma enganosa. entretanto, ficarei devendo essa, porque nunca estudei esse método.

If you have multiple treated units, you can use standard methods like jackknife (e.g. in the sdid paper + package).

Single treated unit case is hard: permutation test / conformal / prediction intervals.

Hi, I guess you noticed the imposter, but as a heads up for @splashzone.bsky.social as well - you have an impersonator (bsky.app/profile/sdid...). Please report and block them appropriately. #SpotTheScammer #ScamAlert

I KEEP NOT LOOKIBG AT FOLLOWING POSTS GOD HOW SDID I MISS PEAK

C’est la même chose dans tous les domaines avec un peu d’expertise (i.e tout ce qui n’est pas du copywriting et du conseil en strat) : les démo IA à base de LLM/SDid impressionnent les technobro, mais les spécialistes du domaine baillent tellement c’est convenu, superficiel, sans substance…

Likewise, looking at our own findings on anti-LGBTQ resolutions in Poland (Haas et al., forthcoming in the JOP), separate-weights SDID estimates yield a sum of 0.77pp. With common weights, the sum constraint holds, the demobilization effect shrinks, but the incumbent gain loses significance.

There's an issue with the caption. The employment numbers are seasonally adjusted, while the hours data is not (BLS does not publish SA hours data at the state level). The SDID method should help absorb some of the seasonality and our robustness checks confirm that this is not a MN winter effect.

We implement common weights that jointly minimize the sum of squared pre-treatment deviations between the weighted controls and the treated outcomes, across all components. Our R package (github.com/lstoetze/pro...) extends synthdid, adding multi-outcome support for SDID and the classic SC method.

My hypothesis is if we applied same SDID method @johnspringford.bsky.social & I used for Brexit migration impacts to the normalisation of racist/far right political discourse we'd find no/small impact.

@robfordmancs.bsky.social @profsob.bsky.social
@tabouchadi.bsky.social

Want to try?

He theorizes how institutions, elections, & race constrain legislators of color and affect policymaking. He is the only scholar, to my knowledge, using advanced causal inference methods (SDID & synthetic control) to answer questions about legislatures, race & representation. Quoting his diss.: