I'm so happy to announce version 2.0.0 of my #Rstats package WeightIt is out on CRAN! New features: censoring weights, multilevel propensity scores, improved weights for continuous treatments, bias-reduced ordinal and multinomial models, M-estimation in subgroups Check out the website below!

Weighting for Covariate Balance in Observational Studies

Generates balancing weights for causal effect estimation in observational studies with binary, multi-category, or continuous point or longitudinal treatments by easing and extending the functionality ...

ngreifer.github.io

Making a causal claim, however weak and tentative, without even listing all your assumptions = Asking for the million dollar lottery prize without demonstrating that you have all matching numbers

New blog post! Odds ratios are...a bit weird. Here I explain two problems (interpretation, non-collapsibility) and provide some simple recommendations for improved reporting. www.the100.ci/2026/07/30/r...

Reviewer notes: Odds ratios are really odd

When psychological researchers investigate binary outcomes, they routinely report odds ratios to quantify the strength of an effect or an association. This is quite understandable given that psycholog...

the100.ci

Saying your results are enough evidence for doing an RCT is a causal claim. If you didn't use causal inference to arrive at that claim, you failed to use the very tool designed to answer your question.

I use this as my primary, non-academic emails and it's great. Plus heat from the servers is used to heat houses.

Infomaniak@infomaniak.bsky.social · 7d ago

"Just" a webmail service developed, hosted and secured in Switzerland, protected by one of the strictest data management laws in the world. 1.2 million logins a day. And if you haven't yet, now's the time to join us 👉 https://www.infomaniak.com/gtl/ksuite.service-mail

A new paper on bounds for per-protocol effects in trials My motivation in writing this paper was to twofold: (1) provide a valid way to estimate per-protocol effects and (2) make the computational procedure as simple as possible to apply link.springer.com/article/10.1...

Computing assumption-lean bounds for per-protocol effects - Trials

Background A common approach used to estimate per-protocol effects, or the effect of perfect adherence to a defined protocol on a specified outcome, is to artificially censor individuals when they dev...

link.springer.com

Is it useful to measure “biological” v “chronological” age? [Spoiler: Nah❗] www.economist.com/science-and-... At the individual level, biological age tests mostly nonsense. It can also lead to harmful “biomarker chasing” (obsessing on wellness longevity BS). Advice ALWAYS same: #healthylifestyle

Is it useful to measure “biological” v “chronological” age?

You will get data, but perhaps not knowledge

economist.com

There is a major problem with how applied health (equity) researchers think about covariate control. I often read research papers where the paper is trying to answer an important question....But the adjustment set doesn't make sense. #episky 1/3

"A Framework for Thinking About the Potential Public Health Impact of Epidemiologic Research". #EpiSky www.ovid.com/jnls/epidem/...

A Framework for Thinking About the Potential... : Epidemiology

An impactful epidemiologic question is one that, if answered, could inform meaningful action to reduce the burden of disease in the population it...

ovid.com

It's the "I'm just asking questions" of observational research. I'm just sharing my results! People can interpret as they see fit!! Never mind that you need several years of epi training to identify the problems with your analysis (which may or may not be gestured at in three lines of limitations).

I'm looking to develop some (internal lab group) guidelines for lab notebooks for epi/biostats/stats projects. Does anyone have good resources on this? There is lots of wet-lab stuff, but I have seen less for epi research (but I think that record keeping is important to do)

I recently wrote about how applied causal inference can align with what Feynman called "cargo cult science"; namely, when we state identification assumptions without interrogating them. I use an example from nutritional epidemiology to convey the problem: miceandtigers.substack.com/p/cargo-cult...

Cargo Cult Causal Inference

Identification is a proof; real-data analysis is an inverse problem. Reciting causal assumptions instead of interrogating them is cargo cult causal inference.

miceandtigers.substack.com

I often come across: expressing the effect estimate in terms of SD makes it more interpretable. But I rarely find that this is actually the case unless we're talking about a latent variable or some scale that is really uninterpretable.

In this day and age, the first two comments by the reviewers should disqualify their reviews.

Georgia Tomova@georgiatomova.bsky.social · last mo.

Just got the following "advice" from the panel regarding my rejected funding application for a causal inference study in observational data 🙂🔫 - present results as correlational rather than causal - avoid causal language - recognise confounding and selection ...and more. It's the full bingo card!

Made it to Phoenix 🌵☀️ & ready to kick off #SER2026 with a session on peer review for epi ECRs tomorrow (Tues) at 8:30AM with the fabulous @tfeend.bsky.social!

Yeah, it kinda stinks

If you can't attend the ABC's workshop, all the materials are available online: github.com/pzivich/ABCs...

GitHub - pzivich/ABCs_of_M-estimation

Contribute to pzivich/ABCs_of_M-estimation development by creating an account on GitHub.

github.com

Causal inference with data that is MC (missing completely)

Adam L@adam-lg.bsky.social · 2mo ago

This is a very cool looking paper, but the title ("Causal Inference with Missing Exposures and Missing Outcomes") is funny to me: it sounds like you can estimate treatment effects when you're missing both exposures and outcomes for the same observation. Paper: arxiv.org/abs/2506.03336 #statssky

Hockey Night in Canada is quintessential Canadiana. I go into one hearing for 8 ½ hours and I come out and Carney loses Hockey Night in Canada 🇨🇦? No.

Robeyns in the Guardian: www.theguardian.com/commentisfre... (for the trillionth time)

Is it bad that Elon Musk has a trillion dollars? Yes, and here’s why | Ingrid Robeyns

Just as the ‘poverty line’ determines what’s required for basic living, we need a ‘wealth line’ to show when extreme wealth becomes harmful, says economist and philosopher Ingrid Robeyns

theguardian.com

Just say no to gift authorship.

Atheendar Venkataramani@atheendar.bsky.social · 2mo ago

I'm reaching that stage of my #academic #medicine career where papers somehow "just materialize." I'm trying to run away from it as fast as possible!!

I don’t know much of the context or backstory here but this seems to me like a model example of how to engage with online critique So rare to see it makes you stop and do a double take

Aaron Moss@aaronjmoss.bsky.social · 2mo ago

Hey Malte, my co authors and I saw your review on PubPeer and we're gratified that your group engaged so thoughtfully with the paper. We'd be happy to join your journal club in the future to respond, answer questions you have, etc. Let me know if you're interested

it’s disgusting that UCSF welcomed back vinay prasad with open arms to teach public health after what he did at the FDA

Like every research tool, the problem is not the tool but the human. LLMs are compatible with whatever kind of science *you* believe in and practice. If you believe in being patient and self-critical, LLMs can help you do that. If you believe in rushing to publish more beans, they do that too.

Georgia Tomova@georgiatomova.bsky.social · 2mo ago

Are LLMs ever compatible with slow science? Even if they aren’t used to necessarily “produce more”, don’t they remove the friction that would normally make us stop, think, and slowly figure things out? And isn’t slow science mostly about taking the time to think?

Re-upping for the Monday morning crowd.

Jeremy Labrecque@jeremylabrecque.bsky.social · 2mo ago

Interested in doing a PhD with me in epidemiology on causal inference and triangulation? In beautiful Rotterdam no less. www.werkenbijerasmusmc.nl/en/vacancy/1... Pro tip: if you get to the interview stage, there WILL be a question about ice hockey

Interested in doing a PhD with me in epidemiology on causal inference and triangulation? In beautiful Rotterdam no less. www.werkenbijerasmusmc.nl/en/vacancy/1... Pro tip: if you get to the interview stage, there WILL be a question about ice hockey

Vacature: PhD student in epidemiology to develop and apply causal triangulation

Many key questions in epidemiology are fundamentally causal. Yet, we often have to rely on assumptions that cannot be directly verified to judge whether an observed association truly reflects causalit...

werkenbijerasmusmc.nl

Things that you didn't know were AI: The Dewey decimal system Stata's merge command Clippy Ctrl + Z

Kalshi now allows you to place Pascal’s Wager

Is novelty really a “key ingredient” for high quality research? To me quality and novelty are independent qualities.

Science Magazine@science.org · 2mo ago

Novelty is often called a key ingredient for high-quality research, but it is notoriously thorny to define or measure. A new tool that puts numbers to how novel a paper is could help—and might be used to more efficiently evaluate the growing flood of scientific manuscripts. https://scim.ag/4et5e3I