1. The paper with the implausibly large effects of Omega-3 fatty acids on mental health was now retracted. A little thread on the process where @ianhussey.mmmdata.io and I was involved.
www.sciencedirect.com/science/arti...

How AI use in scholarly publishing threatens research integrity, lessens trust, and invites misinformation Since 2023, a significant number of published scholarly papers show signs of having been edited using AI tools. These tools are also being used to review papers and search and discovery tools, in…

How AI use in scholarly publishing threatens research integrity, lessens trust, and invites misinformation

48th #Retraction for Didier Raoult - by @plosone.org - because the IRB approvals did not cover all regions and experiments.
PubPeer concerns raised by me in March 2021: pubpeer.com/publications...
journals.plos.org/plosone/arti...

When the Mean is Misleading: a Guide to Ordered Regression for Meaningfully Modeling Ordinal Outcomes - Journal of Quantitative Criminology Journal of Quantitative Criminology - We demonstrate the utility of ordered regression models for analyzing ordinal outcomes frequently encountered in criminology, comparing their predictive...

Just saw this new paper by Brauer on how linear models of mean effect can be very misleading for ordinal data link.springer.com/article/10.1...

The graphical representation of clinical trials with particular reference to measurements over time - PubMed Clinical trials are frequently analysed as if they were surveys when they are in fact experiments. In particular the experimental basis of clinical trials is rarely reflected in the graphs which are used to illustrate them. Proposals are made, and illustrated by example, as to how appropriate figure …

So you carried out a controlled clinical trial? Bravo. Make sure you carry out a controlled analysis.
You can bin those within-group standard errors, comparisons back to baselines and dynamite detonator plots for a start.

pubmed.ncbi.nlm.nih.gov/2277879/

Nice paper on effective sample size for Kaplan-Meier survival curve estimates: www.tandfonline.com/doi/full/10.... #Statistics #StatsSky

SITAR-d: extending the SITAR growth curve model to allow for variability in post-pubertal velocity The SITAR growth curve model with its three random effects is useful for summarising height growth in individuals and groups. However SITAR performs less well for weight which unlike height continu...

𝗦𝗜𝗧𝗔𝗥-𝗱: 𝗲𝘅𝘁𝗲𝗻𝗱𝗶𝗻𝗴 𝘁𝗵𝗲 𝗦𝗜𝗧𝗔𝗥 𝗴𝗿𝗼𝘄𝘁𝗵 𝗰𝘂𝗿𝘃𝗲 𝗺𝗼𝗱𝗲𝗹 𝘁𝗼 𝗮𝗹𝗹𝗼𝘄 𝗳𝗼𝗿 𝘃𝗮𝗿𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗶𝗻 𝗽𝗼𝘀𝘁-𝗽𝘂𝗯𝗲𝗿𝘁𝗮𝗹 𝘃𝗲𝗹𝗼𝗰𝗶𝘁𝘆. Cole, T. J., Sandhu, S., & Elhakeem, A. (2026). @annalshumanbio.bsky.social, 53(1). doi.org/10.1080/0301....

I wrote a perspective for Anesthesia & Analgesia -
"Seeing Is Believing? Scientific Misconduct and the Detection of Problematic Images"

Still trying to make it Open Access. Fingers crossed!

journals.lww.com/anesthesia-a...

%%title%% %%sep%% Maastricht University Press Open Access book by Luc Smits, Sander van Kuijk, and Laure Wynants: This open-access textbook offers a practical and comprehensive guide to developing, validating, and implementing clinical prediction...

✨ New open-access book✨
I’ve seen how powerful prediction models can be, but also how often they fall short. We wrote a book, covering not just development, but also when models are needed, and how to ensure real-world impact.
www.maastrichtuniversitypress.nl/cpm

Signal or noise? Evaluating commonly used attribution methods for explaining deep neural networks in electrocardiogram classification AbstractAims. Attribution-based explainability methods are widely used in electrocardiogram (ECG) analysis to interpret predictions from ‘black-box’ deep n

Fantastic new paper casting doubt on explainability of explainable AI. To explain complex machine learning algorithms you need reproducibility of the explanation at a minimum academic.oup.com/ehjdh/articl... #machinelearning #Statistics #StatsSky @maartenvsmeden.bsky.social

A comprehensive public health approach is needed to study the impact of digital technology on health Barely a day goes by without media and political concern about the impact of digital technology on health, particularly among young people. The publication of a recent government report, “Understandin...

"Research should not only be independent of vested interests but must also be seen to be so"

My @bmj.com piece with @fphuk.bsky.social president Tracy Daszkiewicz argues we need a public health approach to study the impact of digital technology on health.

#CDOH

www.bmj.com/content/392/...

Online Studies Psychological Science requires that authors who use samples from online data collection include a statement in the Method section explicitly addressing their approach to preventing and detecting automated or AI-generated responses. Rationale As large language models and other generative AI tools become more accessible, the risk of data contamination by non-human respondents has increased dramatically in research. Psychological science (and the social sciences generally) is particularly susceptible to this issue given its growing reliance on online data collection. Preventing automated responses during data collection and detecting them afterward often involve methodological trade-offs. For instance, technical barriers that aim to prevent LLM use (e.g., blocking copy-pasting functionalities) may eliminate behavioral indicators needed for detection (e.g., pasting rather than typing). This policy aims to enhance transparency and reproducibility of reported results by requiring authors to articulate their approach across both prevention and detection dimensions, enabling readers and reviewers to assess the likelihood of reported data being influenced by automated responses. Scope This policy applies to any submission with at least one study that includes data collected online without direct human supervision (e.g., via crowdsourcing platforms, student participants who complete the study online, online recruitment ads, or remote survey distribution tools). Required Reporting Authors must include in the Methods section either: A statement confirming that procedures were in place to prevent and/or detect and exclude automated or AI-generated responses, including a description of those procedures (e.g., explicit participant instructions against LLM use, disabled copy–paste functionality, CAPTCHA use, IP filtering, consistency checks, attention checks, adversarial prompting) as well as the types of automated responses that these procedures are suitable …

Maybe of interest: The submission guidelines of Psychological Science now demand an explicit statement on measures taken to reduce the risk of AI-generated responses for all online studies!

www.psychologicalscience.org/publications...

Ratios are misleading exposure variables that compromise a regression model unless fundamental scaling assumptions are satisfied: a comment on Valente et al - Journal of Clinical Epidemiology www.jclinepi.com/article/S089...

An Empirical Assessment of the Cost of Dichotomization of the Outcome of Clinical Trials We have studied 21 435 unique randomized controlled trials (RCTs) from the Cochrane Database of Systematic Reviews (CDSR). Of these trials, 7224 (34%) have a continuous (numerical) outcome and 14 211 (66%) have a binary outcome. We find that trials ...

Want to make your research more meaningful? Ornithology? Smear the lenses of your binoculars with vaseline. You may not be able to tell a hawk from an eagle but you won’t mistake a sparrow for an ostrich.
Clinical research? Why not use responder dichotomies?

pmc.ncbi.nlm.nih.gov/articles/PMC...

for critique see here:
pubpeer.com/publications...

Improved Centile Estimation by Transformation And/Or Adaptive Smoothing of the Explanatory Variable A popular approach to growth reference centile estimation is the LMS (Lambda-Mu-Sigma) method, which assumes a parametric distribution for response variable Y$$ Y $$ and fits the location, scale and ....

𝙄𝙢𝙥𝙧𝙤𝙫𝙚𝙙 𝘾𝙚𝙣𝙩𝙞𝙡𝙚 𝙀𝙨𝙩𝙞𝙢𝙖𝙩𝙞𝙤𝙣 𝙗𝙮 𝙏𝙧𝙖𝙣𝙨𝙛𝙤𝙧𝙢𝙖𝙩𝙞𝙤𝙣 𝘼𝙣𝙙/𝙊𝙧 𝘼𝙙𝙖𝙥𝙩𝙞𝙫𝙚 𝙎𝙢𝙤𝙤𝙩𝙝𝙞𝙣𝙜 𝙤𝙛 𝙩𝙝𝙚 𝙀𝙭𝙥𝙡𝙖𝙣𝙖𝙩𝙤𝙧𝙮 𝙑𝙖𝙧𝙞𝙖𝙗𝙡𝙚. Rigby RA, Stasinopoulos DM, Cole TJ. Stat Med. 2026 Feb;45(3-5):e70414. doi: 10.1002/sim.70414.

onlinelibrary.wiley.com/doi/10.1002/...

Fascinating stuff. However, IMO, the value of double blind trials is not just dealing with patient- centred biases but physician-biases and this is often overlooked. www.linkedin.com/pulse/blind-...?

Die, Dichotomy We have studied 21 435 unique randomized controlled trials (RCTs) from the Cochrane Database of Systematic Reviews (CDSR).Of these trials, 7224 (34%) have a continuous (numerical) outcome and 14 211 (...

Die, Dichotomy.
www.linkedin.com/pulse/die-di...
A brief post encouraging you to read the original paper with @erik-van-zwet.bsky.social and @f2harrell.bsky.social .
onlinelibrary.wiley.com/doi/epdf/10....

How to Do Bad Biomarker Research – Statistical Thinking This article covers some of the bad statistical practices that have crept into biomarker research, including setting the bar too low for demonstrating that biomarker information is new, believing that...

Judging by the poor quality of biomarker research I see reported in biomedical journals, my article "How to Do Bad Biomarker Research" must have been hugely influential: www.fharrell.com/post/badb/in... #Statistics #StatsSky #EpiSky

Editor in Chief publishes eight articles in one issue of his (Scopus indexed) journal. Is that ethical?
🧵

“You will find that many alternative medicine enthusiasts are into proidiotics.” Sayings of Confuseus

“Assessing the properties of the prediction interval in random-effects meta-analysis” www.cambridge.org/core/journal...

Our letter to editor about Caldwell et al (J Appl Physiol 139: 1220–1227, 2025) is published: journals.physiology.org/doi/full/10....

There seemed misconceptions about variance comparison stats. We also commented on the author's “SD of treatment effects”. Suppl stuff: zenodo.org/records/1777...

#neuroscience #biomedicine #researchintegrity

How long should it take to retract a paper with incontrovertible signs of data fabrication? Sleuths think 2 months is too long, particularly when clinical risks are involved. deevybee.blogspot.com/2026/01/an-o...
#retraction #stemcells #cardiology
@erictopol.bsky.social

Building an AI Scientist Hertz Fellow Sam Rodriguez launched FutureHouse, a nonprofit research lab working toward building an AI scientist or AI systems that can automate scientific research in biology and other complex scien...

“An AI scientist, for example, could figure out how the human brain works, or deliver any gene to any cell in the body.”

Yeah and a magic pony could shit bricks of 24k gold and piss a highly concentrated solution of pure heroin.

Most researchers would receive more recognition if assessed by article-level metrics than by journal-level metrics buff.ly/EhQn7JQ

Figure 1: The Rothman-Dahly Evidence Pyramid (original version) An equilateral triangle with a small blue section labelled "Thoughtful, well-conducted studies of any design" at the top, with the remaining space colored red and labelled "The other shit"

‪It has a name now 😜

Many thanks to Ken for agreeing to put his good name to my...artwork. The image is in the public domain (CC 0), but citations to the linked documents are warmly welcomed.

✅ zenodo.org/records/1808...

✅ pubmed.ncbi.nlm.nih.gov/24452418/

The best bit about the show Tipping Point is when the host, Ben Shephard, says "let's find out what would have happened if you'd decided to play". There's always a nagging doubt at the back of my mind: that's not really what *would* have happened, is it?

“Re-examination of the 3/4-law of metabolism” and “Toward a metabolic theory of ecology”
statmodeling.stat.columbia.edu/2025/12/18/r...