Tilman Bayer · Jul 2

See also this thread (about the entire article) bsky.app/profile/ding...

Jarrett Byrnes · Jul 1

Tim Triche, Jr.🔰 · Jul 1

Dear god

Marcus Crede · Jul 1

Ryan Hisner · Jul 2

Is the problem in assuming that including more covariates are always better? Or in asserting that observational evidence + theoretical considerations are often the best option? I'm too dull to tell.

Tristan Snowsill · Jul 1

I just love the idea that confounders are floating around there, readily measured, and it's really just a birthday paradox problem or something.

Seungju Kim (he/him) · Jul 2

Im sure there are many reasons why this is problematic, at least in clinical psychology, it’s a measurement problem for me. Are there others (more consequential) that I am missing?

Thom Volker · Jul 1

At the risk of making an idiot of myself in public: most of this is not wrong, right? I cannot really comment on the statement that there are usually too few covariates in ph, but it is strictly more likely that you can block the paths that need to be blocked if you have access to more variables.

Rafael Pinto · Jul 2

@jhfflores.bsky.social

Mathieu M.J.E. Rebeaud · Jul 1

I've heard @statsepi.bsky.social shouting from here...

Pat Savage · Jul 1

Mad (Social) Scientist 🇨🇦 · Jul 2

We call those 'garbage can' models in my neck of the woods. (Which, admittedly, isn't public health, but I'd imagine some of the same considerations apply.)

Casaubon · Jul 1

Lord knows I took far too many metrics classes to not feel like a troglodyte for asking this but…what specifically is wrong with this? I’m forgetting so much.

Thomas Kealy · Jul 1

That paragraph is literally painful for me to read.

Rafe Meager (they/them) · Jul 1

Ben Hanowell · Jul 2

This reminds me of a conversation I had with the a principal applied scientist at a certain large firm I once worked at. Guy legitimately thought the best way to do causal inference was to throw as many variables in autoML as we could then do SHAP.

@kfan99.bsky.social · Jul 2

Hey Johannes This isn't just another post in your feed. It's a family's story, shared with hope that someone will care. Please visit my page and read the pinned post. ❤️

Saloni · Jul 1

From open.substack.com/pub/asterisk... Side noting that I think the author may not know you can also adjust for covariates in an RCT...

Lachlan Cribb · Jul 2

Besides the last statement, none of this seems unreasonable? A richer dataset will make it easier to adjust for confounding, all else equal. And most PH research probably does suffer more from underadjustment than from overfitting.

Darren Dahly · Jul 1

I'm afraid this person is waaaay too early into the their career for me to comment ;)