Causal diagrams and counterfactual reasoning reach political sciences. Congratulations to Prof Jacobs UBC Political Science on the successful launch of his book!
Cameron Patrick These lectures by Brady Neal youtube.com/playlist?list=… or these by Boris Sobolev youtube.com/playlist?list=…
The primer is also very useful amazon.com/Causal-Inferen…
On May 16 hear from UBC's School of Population and Public Health PhD candidate Sean on his coronary revascularization study comparing patient outcomes for those receiving delayed #CABG vs timely #PCI . Details & registration (req'd): chspr.ubc.ca/2023/04/21/chs… C2E2 VCH Research Institute Boris Sobolev Mike Law
Stephen John Senn Elias Bareinboim Boris Sobolev Another ideas from BOW are not provocative... are simply wrong and dangerous! This pieace is from a paper of James Robins (another observational folk who promotes OS in medicine with falsehoods).
Matej Zečević Lianhui Qin Amit Sharma Microsoft Research Devendra Singh Dhami | देवेंद्र सिंह धामी Kristian Kersting Emre Kıcıman Judea Pearl Guy Van den Broeck Andrew Lampinen emily mcmilin Kai-Hendrik Cohrs Aniket Vashishtha Oscar Clivio Alessandro Palmarini Melanie Mitchell Angelika Romanou And for the dessert, I met Boris Sobolev for a tea and a discussion on challenges and opportunities in causal modeling in healthcare.
This was after also meeting Scott Mueller earlier that day.
What a ride!
Thank you to everyone who supported us and shared their ideas 🙏🏼
18/18
The most important of Rubin's assumptions here is ignorability: if the untreated were treated, they would have the same outcomes as the treated.
This was, of course, brilliant at that time.
Judea Pearl got rid of the ignorability assumption by proving that ignorability follows
In medical research, a 'good statistician' is one who does exactly as they are asked & is able to produce 'a significant p-value'.
If, instead, you dare to ask difficult questions or point out flaws then you will be branded 'difficult to work with' and get exiled!
#EpiTwitter
This school of stats practitioners continues to equate causal effect with the output of statistical analysis of RCTs, without explicitly specifying the estimand for their estimators.
In contrast, Judea Pearl first defines the causal effect as a target parameter of the
'Some papers' is just a sleight of hand by Frank Harrell. Badly done science doesn't invalidate the scientific method.
If someone can’t perform Rach-4, it doesn’t mean that Rachmaninoff wrote bad music.
Judea Pearl Frank Harrell Boris Sobolev Stephen John Senn Peter Tennant PATE, which is the expected difference between counterfactual outcomes in the target population = 𝔼[Y(1)−Y(0)]
SATE is referring to the sample under study: \frac{1}{n} \sum_{i=1}^{n} \left[ Y_{i}(1) - Y_{i}(0) \right]
(or see attached image)
onlinelibrary.wiley.com/doi/10.1002/si…
![@yudapearl @f2harrell @soboleffspaces @stephensenn @PWGTennant PATE, which is the expected difference between counterfactual outcomes in the target population = 𝔼[Y(1)−Y(0)]
SATE is referring to the sample under study: \frac{1}{n} \sum_{i=1}^{n} \left[ Y_{i}(1) - Y_{i}(0) \right]
(or see attached image)
onlinelibrary.wiley.com/doi/10.1002/si…](https://pbs.twimg.com/media/GOWJgQ0WQAEQO2V.png "Tim Kaiser (@tkaiser_science) on Twitter photo 2024-05-24 13:11:30 @yudapearl @f2harrell @soboleffspaces @stephensenn @PWGTennant PATE, which is the expected difference between counterfactual outcomes in the target population = 𝔼[Y(1)−Y(0)]
SATE is referring to the sample under study: \frac{1}{n} \sum_{i=1}^{n} \left[ Y_{i}(1) - Y_{i}(0) \right]
(or see attached image)
onlinelibrary.wiley.com/doi/10.1002/si…")
Murat Kocaoglu Boris Sobolev Hi Murat, thanks for your note and congrats again on your work! Having said that, I feel the comparison you make is a bit misleading and misreads the real technical nature of R-80 and its contributions. In case you didn't have the time to read the paper (including the
