Profile avatar
jeremylabrecque.bsky.social
Canadian epidemiologist and causal inference person at Erasmus Medical Center. Big fan of Northern Expsoure and Car Talk. jeremylabrecque.org
653 posts 2,277 followers 1,138 following
Regular Contributor
Active Commenter
Posts 16 Comments 34

Finally, something AI can do just as well as real "researchers". www.nature.com/articles/d41...

submitted 1 day ago • 3 comments

Truths by Stephanie: Precision in language is useless if people don't/won't engage for the time required to understand what it says.

submitted 6 days ago • 1 comment

Null hypothesis significance testing makes people say the craziest things. "The was no difference in risk between exposed and unexposed (risk difference: 10%, CI: -10%,30%)" That people can say "no difference in risk" then "the difference in risk was 10%" right next to each other blows my mind.

submitted 5 days ago • 6 comments

I would love to think that everyone has known for a long time that LLMs aren't going to lead to "general intelligence," but all the evidence suggests there are plenty of people who resist this, so it's good to have a paper from researchers Apple making the point strongly.

submitted 9 days ago • 14 comments

Is there a way to depict positivity (or a lack thereof) on a DAG? A reviewer asked for this and I can't think of how you would show this.

submitted 7 days ago • 8 comments

It's #MEstimatorMonday (23/52) and #SER2025 eve (so don't forget to leave coffee out for the epidemiologists) Today's M-estimator comes from a student-led paper and a good next stepping stone for those attending the ABC's of Estimating Equations workshop tmrw pubmed.ncbi.nlm.nih.gov/40167173/

submitted 9 days ago • 2 comments

The French also respond “Please don’t do that” when you speak to them in québécois French.

submitted 9 days ago • 0 comments

Starting to look like I might not be able to work at Harvard anymore due to recent funding cuts. If you know of any open statistical consulting positions that support remote work or are NYC-based, please reach out! 😅

submitted 13 days ago • 11 comments

Have you wondered what an artificial neural network really does (and how it might just be regression all the way down)? Well I have a new AJE Classroom paper on the basics of how neural networks work academic.oup.com/aje/article/...

submitted 14 days ago • 5 comments

Eso po

submitted 15 days ago • 1 comment

To coincide with a new paper in Epidemiology from our group, let's talk about the connection with influence functions for #MEstimatorMonday (22/52) journals.lww.com/epidem/fullt...

submitted 16 days ago • 1 comment

DON'T MISS: "Complexities & Advantages of Causal Methods for Social Questions: An applied target trial emulation study of migration & mortality in South Africa" The next #CIIG seminar with Rachel Yorlets on 2nd June 2025 at 3pm BST! Register: turing-uk.zoom.us/meeting/regi... #EpiSky #CausalSky

submitted 26 days ago • 0 comments

CEO pay is up 1,085% since 1978, while typical worker pay is up just 24%.  Why do we always hear "we can't afford to pay our workers more" but never "we can't afford to pay our CEO more"?

submitted 25 days ago • 517 comments

I can't stop thinking about this figure.

submitted 26 days ago • 10 comments

In epi, the desire to control confounding is so pathological that people always assume that that's the goal of every method. So often I see "we used IV analysis to control for confounding." IV doesn't control confounding, it just doesn't care about (exposure-outcome) confounding.

submitted 27 days ago • 0 comments

I haven't been advertising well with #MEstimatorMonday but there is an M-estimator (estimating equations) workshop at SER this year and you can still sign up We will cover like the first 10ish Mondays in more detail. It would be a good setup for the rest of the year epiresearch.org/annual-meeti...

submitted 28 days ago • 1 comment