Jeremy Labrecque
banner
jeremylabrecque.bsky.social
Jeremy Labrecque
@jeremylabrecque.bsky.social
Canadian epidemiologist and causal inference person at Erasmus Medical Center. Big fan of Northern Expsoure and Car Talk.

jeremylabrecque.org
Probably 10 times a day (at least) I run into some form of this "not significant therefore null is true" problem.

There were many, many problems in the undergrad stats course I took, but one positive is they repeated a million "you cannot accept the null hypothesis."
November 3, 2025 at 9:44 AM
Reposted by Jeremy Labrecque
🚨Wow - this will be huge if it bears out. Exit polls show the pro-EU D66 party has WON the Dutch election.

The far-right PVV party of Geeet Wildsrs has lost 12 seats, according to exit polls.

Huge if true.
October 29, 2025 at 8:09 PM
Just read a paper that looked for "interactions" of the main exposure with like five different variables. Which me realize that this is basically the same problem as "risk factor" analyses except with interactions instead of main effects.
October 29, 2025 at 12:59 PM
Qobuz
Spotify is now running ICE recruitment ads. We asked them to stop. They ignored us. Let's show them what we showed Disney. No Kings, No Collaborators, No Capitulators. indivisible.org/cancel-spotify
Don’t Stream Fascism: Cancel Spotify
indivisible.org
October 29, 2025 at 12:47 PM
Reposted by Jeremy Labrecque
Factors associated with: problems of using exploratory multivariable regression to identify causal risk factors bmjmedicine.bmj.com/content/4/1/...

I know who'll enjoy reading this: @jeremylabrecque.bsky.social
Factors associated with: problems of using exploratory multivariable regression to identify causal risk factors
Many medical and epidemiological studies use multivariable regression to test whether several independent variables (exposures) are causal determinants of a health outcome. Where mutually adjusted reg...
bmjmedicine.bmj.com
October 28, 2025 at 11:34 AM
Reposted by Jeremy Labrecque
Chatbots — LLMs — do not know facts and are not designed to be able to accurately answer factual questions. They are designed to find and mimic patterns of words, probabilistically. When they’re “right” it’s because correct things are often written down, so those patterns are frequent. That’s all.
June 19, 2025 at 11:21 AM
In the past few months, I've stopped supporting all major US corporations (e.g. subscriptions). I thought it was going to be annoying but instead I've discovered so many better options like Qobuz for music streaming (which pays artists 4x more) and Infomaniak for file storage. Pure wins.
October 27, 2025 at 4:45 PM
Reposted by Jeremy Labrecque
POV: The new singer of the Reading Rainbow theme song and the new host of Reading Rainbow meet for the first time in Canada 🥹📚🌈

Video Description:
Mychal and Bukola meet for the first time! They joyfully dance to the Reading Rainbow theme song, sung by Bukola ✨

youtube.com/@kidzuko?feature=shared
October 25, 2025 at 6:25 PM
Reposted by Jeremy Labrecque
Are there earlier occurrences of groups being established to archive biological data to make them available for future analyses by researchers than this, which is from The British Association Committee on Biological Measurements, 1927?
October 24, 2025 at 10:45 AM
Reposted by Jeremy Labrecque
Great piece on the absurdity of brute force multiverse analyses.

www.pnas.org/doi/10.1073/...
Robustness is better assessed with a few thoughtful models than with billions of regressions | PNAS
Robustness is better assessed with a few thoughtful models than with billions of regressions
www.pnas.org
October 22, 2025 at 5:29 PM
Reposted by Jeremy Labrecque
Research is mostly discovering cleaner ways to express what you overcomplicated yesterday
October 22, 2025 at 9:12 AM
Reposted by Jeremy Labrecque
The recording is now available so that you can confirm that I indeed have a German accent and color-match my outfits with my Zoom background.

youtu.be/YL0co26ng-g?...
October 21, 2025 at 3:16 PM
Good thing it doesn’t fall near April 27 which is King’s Day here in the Netherlands. That would be awkward
Lol that the London No Kings event is called No Tyrants because…you know…

www.nokings.org#map
October 16, 2025 at 9:17 PM
Reposted by Jeremy Labrecque
Lol that the London No Kings event is called No Tyrants because…you know…

www.nokings.org#map
October 16, 2025 at 7:32 PM
Reposted by Jeremy Labrecque
Most graphs of the fertility rate depict the 'period fertility rate', which is based on a single year's data and doesn't necessarily reflect how many children women actually have across their lifetimes.

I've used data from the Human Fertility Database to show the cumulative number instead:
October 1, 2025 at 9:36 AM
Reposted by Jeremy Labrecque
Science is grounded in observation. Measurement is a tool for observation. Measurements should be evaluated for validity and reliability/uncertainty. Scientists who use measurements without understanding their properties are not really scientists at all.
Can large language models stand in for human participants?
Many social scientists seem to think so, and are already using "silicon samples" in research.

One problem: depending on the analytic decisions made, you can basically get these samples to show any effect you want.

THREAD 🧵
The threat of analytic flexibility in using large language models to simulate human data: A call to attention
Social scientists are now using large language models to create "silicon samples" - synthetic datasets intended to stand in for human respondents, aimed at revolutionising human subjects research. How...
arxiv.org
October 1, 2025 at 5:39 AM
Reposted by Jeremy Labrecque
Since the Ezra & Ta-Nehisi discussion is still happening: the main point I think most are missing is that Klein is saying the role of the journalist-intellectual is to do strategic politics, whereas Coates says the role of the journalist-intellectual is to tell the truth
September 30, 2025 at 2:28 PM
One reason I oppose associational or other non-causal language in research questions is it lets people think they don't need causal inference to answer causal questions.

(When the underlying question is causal, of course. Which it almost always is.)
September 30, 2025 at 7:52 AM
Reposted by Jeremy Labrecque
I'm gonna be saying "was silence not an option?" from now on
September 29, 2025 at 12:36 AM
Reposted by Jeremy Labrecque
Congrats to Moldova!!! Russia threw everything to destabilize the election and failed spectacularly.

Democracy won 🙌🏼
September 29, 2025 at 12:50 AM
Reposted by Jeremy Labrecque
So exhausting..
September 28, 2025 at 6:41 PM
Reposted by Jeremy Labrecque
And to be fair, many academic journals also ban "causal" words like effect/impact/due to.
How common are blasphemy laws? About 4-in-10 countries have them.
www.pewresearch.org/short-reads/...
September 27, 2025 at 8:52 PM
Reposted by Jeremy Labrecque
Canada Post “lost” 1 billion dollars last year?

How about, “it cost Canadians 1 billion dollars to have a national postal service” which works out to costing about $25 a year per person (population of Canada in 2024 = 40 million). Seems like a pretty reasonable cost to me.
"Canada Post is on track to lose money" Hum. Duh. It cost less than a toonie to send a letter across Canada.

"Canada Post is a service and not a business" was common knowledge until late stage capitalism brain rotted most people into think if it ain't making money for shareholders, it's failing.
September 25, 2025 at 8:35 PM
Reposted by Jeremy Labrecque
This is not because AI is generally useless. It is a tool that has to be carefully tested for possible uses and implemented in ways that create a net benefit. Like all tools. This is not happening.
So it may point to the possibility that a lot of business decision makers are useless.
"No single study on AI in the workplace is going to be definitive, but evidence is mounting that AI is affecting people’s work in the same way it’s affecting everything else: It is making it easier to output low-quality slop that other people then have to wade through."
AI ‘Workslop’ Is Killing Productivity and Making Workers Miserable
AI slop is taking over workplaces. Workers said that they thought of their colleagues who filed low-quality AI work as "less creative, capable, and reliable than they did before receiving the output."
www.404media.co
September 24, 2025 at 2:36 PM