A special day for Cohort Membear from @clscohorts.bsky.social who travelled to Luxembourg to meet its mascot counterpart from @uni.lu ACADI 🧸🦎 Loved talking about causal inference with compositional data to a lovely audience from @liser.lu and LIH. Many thanks @spilleron.bsky.social & @fnr.lu!

How do we incentivise people to publish fewer papers? We need to live in a world where saying 'I have published over 500 papers' is widely understood as a red flag, not as something to brag about.

Thank you for the invitation and for the great discussion!

Recording now available!

I'm looking forward to this! 😊

Happening TODAY at 3pm GMT 👇

Forecast intervals from my time series model be like

"So, are you illuminating a mediation structure, assuming a mediation structure, or testing for a mediation structure?" "Yes"

I guess Rubin choosing "Missing at Random" for the scenario in which it is evident that the data are not missing at random is a close contender.

Have you ever wondered how the transportability of an estimate may be affected by the choice of effect measure? Join us in Nov to hear from @mpiccininni3.bsky.social on why the causal risk ratio might be more stable and transportable across settings. Sign up: ucl.zoom.us/meeting/regi... #CIIG

A certain type of academic can take anything good and turn it into research waste.

🚨Job alert!🚨 We are looking for a 🧮 quant postdoc researcher to join the AMAZING Jay Das-Munshi @jdasmunshi.bsky.social George Ploubidis @georgeploubidis.bsky.social & myself at @kingscsmh.bsky.social in collab with @clscohorts.bsky.social All info: my.corehr.com/pls/kclrecru... Closes 18 Nov!

The constant over- generalization of their data to everyone’s population (not just the UK) is aggravating. Plus, the work is just a bunch of correlation studies that they treat as causal with, as you note, a huge selection bias. Why isn’t this discussed more in the UK?

It is not everyday that 2 out of the 5 'most read' papers in @intjepidemiol.bsky.social are yours! Rather pleased, especially given both of these took *4-5 years* each from start to publication! Slow science sometimes pays off.

Stop interpreting clinical trials as if they were surveys. www.linkedin.com/pulse/illegi...

arXiv will no longer accept review articles and position papers unless they have been accepted at a journal or a conference and complete successful peer review. This is due to being overwhelmed by a hundreds of AI generated papers a month. Yet another open submission process killed by LLMs.

Interested in learning about handling survey mode effects? Still not too late to register for this online event next week 👇

absolutely breaking news from the BBC

DON'T MISS: "An overlooked stability property of the risk ratio & its practical implications" with @mpiccininni3.bsky.social at 3pm (UK) on 10 Nov 2025 This #CIIG seminar explains an important reason why we might want to estimate risk ratios after all! Register: tinyurl.com/ciig-marco-piccininni

We are predictably turning genuine PPI into tick boxes and blanket policies that could never be met with available, precious resources, thus creating the "need" to simulate the patient perspective with LLMs. The irony of it all. A solution is to not cram PPI where it isn't needed.

Have you ever wondered how the transportability of an estimate may be affected by the choice of effect measure? Join us in Nov to hear from @mpiccininni3.bsky.social on why the causal risk ratio might be more stable and transportable across settings. Sign up: ucl.zoom.us/meeting/regi... #CIIG

Great piece on the absurdity of brute force multiverse analyses. www.pnas.org/doi/10.1073/...

I’m dressing as the age period cohort problem for Halloween

Everyone splitting hairs over clustered standard errors when they could be doing this

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?...

If any scientists working in public health nutrition and infant development want to stop helping people and start commercializing your evidence-free breakthrough product opportunities, this @wellcometrust.bsky.social funded program looks amazing. Apply today! 👇🚀🦠 wellcomeleap.org/form/

New blogpost on my concerns about Wellcome LEAP new $50m program on autism/microbiome links. deevybee.blogspot.com/2025/10/a-le... #autism #microbiome #biomarkers #diagnostic

"Many fields have come up with their own creative ways of making causal claims while maintaining plausible deniability" - @dingdingpeng.the100.ci summarises the problem before offering her tips on how to help bring rigorous causal inference methods into mainstream research!

Fantastic to see 200+ people attend the CIIG today. Thank you @dingdingpeng.the100.ci for a brilliant talk! 🤩 Anyone with a broad interest in causal inference can subscribe to hear about our future seminars and see past recordings here: cls-data.github.io/CIIG/

Join us at 3pm UK time today! 🎉 Check out our webpage and subscribe to our mailing list: cls-data.github.io/CIIG/

Today is the day 👀

Yes; this is basically my point. It can’t be relied on for something this simple and yet… futurism.com/future-socie...

We did this paper as part of a methods-focused grant that required patient involvement and I honestly found it so valuable that I tried to add similar involvement to all my grant submissions. Not that any of those got funded, but… pubmed.ncbi.nlm.nih.gov/29966732/

the best way to get helpful academic engagement is to ask your question on a weekend

Does anyone have any good examples of PPI for stats methodology? I’ve seen a few but they are all related to a clear applied research question, rather than actual methods

No time for peer review. But asked to make time for the car park satisfaction survey.

1/ I guess I’m ‘The Good Immigrant’: Been here most of my life now, never claimed benefits, paid taxes, worked for the NHS, taught students, done research for public good, volunteered, spoke English fluently. So why do I feel so alienated and hurt by the government’s anti- immigrant rhetoric?

sure, sounds great! (me after a 40-minute battle against my overwhelming desire to say no)

"how to win friends and influence people" but mostly "how to win friends"

The microbiome is an easy target for purveyors of pseudoscience.

Should statistical software that estimates causal effects also tell you the causal assumptions under which that estimate can be interpreted as causal? I don't know but my PhD student Maurice Korf has some thoughts (and software) to get the conversation going: academic.oup.com/ije/article/...

or when Judea Pearl said that every time he reads the Book of Why, he learns something new

This of course was a CLASSIC of the genre!

The best of these Pearlisms was when he revealed he didn't understand that it's called a "linear" model because it's linear in the parameters. Or maybe it was the tough nurse that could make you comply w/ tx IRL. Or the retweeting of vile right wing "migrant invasion" accounts. Hard to tell.

Remember when our feed was 99% stats shitstorms? Those were the days ❤️

If you use survey data collected through mixed modes (f2f, telephone, online), check out this @royalstatsoc.bsky.social event with @rjsilverwood.bsky.social @georgiatomova.bsky.social and others.

We’re in the proper McDonald’s era of science now

I recently learnt about "mode effects". I recommend you do too, especially if you work with data from large cohorts. You can hear @georgiatomova.bsky.social, @rjsilverwood.bsky.social, and others discuss what to do with mode effects next month!

After “fast food” And “fast fashion” Have we entered the era of “fast science”?

We are organising a @royalstatsoc.bsky.social Social Statistics Section online event: Handling survey mode effects Wednesday 12 November 2025, 10.00AM - 12.10PM Full info and booking: rss.org.uk/training-eve...