iiinova — where your digital self keeps evolving

Reposted by Greg Atkinson

Peter Tennant @pwgtennant.bsky.social · 8h

"Is an ultra-processed chocolate cake materially more harmful than a home-made one?" I enjoyed contributing to this response to The Lancet series on ultra-processed foods, which outlines some key limitations in the evidence that 'ultra-processing' is specifically harmful to health.

link.springer.com

Public health policy should be grounded in established evidence that acknowledges uncertainty: a response to The Lancet series on ultra-processed foods - BMC Medicine

Background The recent three-part Lancet series on ultra-processed foods (UPFs) argued that the displacement of traditional diets by UPFs is a major driver of poor diet quality and diet-related chronic...

2216

Reposted by Greg Atkinson

LBC @lbc.co.uk · 07/09/2026

This is @mrjamesob.bsky.social's London.

411539461

Reposted by Greg Atkinson

Significance magazine @significancemag.bsky.social · 24/08/2026

Deluded data science? Why statisticians must be crystal clear about what they're doing, before they do it... Read for free now: significancemagazine.com/deluded-data...

021

Reposted by Greg Atkinson

International Journal of Sport Nutrition and Exercise Metabolism @hk-ijsnem.bsky.social · 14/08/2026

Bossi et al have an exciting review in our September issue titled ‘Quantifying and Comparing Training Load Metrics in Cycling: A Methodology Review’. After just releasing our UCI Special Issue, this feels well placed in the fall issue! 🚲 🔗 doi.org/10.1123/ijsn...

021

Reposted by Greg Atkinson

ArXiv Paperboy (Stat.ME+Econ.EM) @paperposterbot.bsky.social · 17/07/2026

arXiv📈🤖 Multiverse analysis, abdication of responsibility and manufacturing of doubt By Modr\'ak

I argue that multiverse analysis is highly suited to two undesirable uses: abdication of researcher's responsibility for their conclusion and manufacturing of doubt. A review of multiverse analyses published in 2025 provides tentative empirical support that abdication of responsibility is present in the literature and I mention anecdotal evidence that multiverse has been used for manufacturing of doubt about Covid-19 precautions. To mitigate negative effects if multiverse analysis becomes widely used I suggest the community adopts two conventions for evaluating multiverse analyzes: evaluating multiverses by the single worst universe they contain and considering large size of a multiverse as a sign of weakness rather than a praiseworthy achievement.

0139

Reposted by Greg Atkinson

Experimental Physiology @expphysiol.bsky.social · 16/07/2026

In this Reply to the Letter to the Editor, Robin Christensen, Tobias Haugegaard, Janus C. Jakobsen (Bispebjerg and Frederiksberg Hospital, Odense University Hospital, @sdu.dk, Rigshospitalet) discuss why reporting requires more than findings 📖 buff.ly/ElWUbOb

021

Reposted by Greg Atkinson

Frank Harrell @f2harrell.bsky.social · 15/07/2026

Continued use of change scores is tragic&results in meaningless non-covariate-adjusted summary stats being communicated out of context. Researchers are acting as if post vs pre is linear with a slope of 1.0 (and floor and ceiling effects can't happen) when they do that. hbiostat.org/bbr/change

hbiostat.org

14 Transformations, Measuring Change, and Regression to the Mean – Biostatistics for Biomedical Research

082

Reposted by Greg Atkinson

Frank Harrell @f2harrell.bsky.social · 15/07/2026

Another example where smart clinical trialists and clinical trial statisticians are completely missing the boat: dichotomization of outcome measures and so-called "responder analysis".

061

Reposted by Greg Atkinson

Peter Tennant @pwgtennant.bsky.social · 14/07/2026

I cannot believe how many of these RCTs used a change/percent change outcomes. Because the literature is so clear on the topic I honestly thought this issue was resolved! @f2harrell.bsky.social @stephensenn.bsky.social

5284

Reposted by Greg Atkinson

Prof Christina Pagel @chrischirp.bsky.social · 14/07/2026

I find this is a really odd story - why should world class biomedical research improve life expectancy in UK? What is keep life expectancy flat are a range of structural issues - not access to the latest medical treatments. 1/4 news.sky.com/story/world-...

news.sky.com

World-class UK medical research not delivering enough benefits for British people

The country's life sciences industry generated £146.9bn in turnover in 2023/24, but UK life expectancy has fallen, a new study has found.

716567

Reposted by Greg Atkinson

Experimental Physiology @expphysiol.bsky.social · 11/07/2026

In this Editorial, Editors-in-Chief Damian Bailey (of Experimental Physiology) and Kim Barrett (of @jphysiol.bsky.social) recognise excellence in peer review ✍️ 📖 buff.ly/AURLI1i

011

Reposted by Greg Atkinson

International Journal of Sport Nutrition and Exercise Metabolism @hk-ijsnem.bsky.social · 06/07/2026

Recognise this logo? Well, there’s a new and improved version of Narang et al’s Time Series Response Analyser for transparent summary statistics from discrete time-series data! 📈📊

tsra-v2.streamlit.app

TSRA v2.0

Time-Series Response Analysis (TSRA) v2.0 is a Streamlit-based research tool for structured analy...

111

Reposted by Greg Atkinson

ArXiv Paperboy (Stat.ME+Econ.EM) @paperposterbot.bsky.social · 29/06/2026

arXiv📈🤖 Experimental Design When N Equals One By Guo, Liang

N-of-1 trials, or time-series experiments, are widely used in clinical research and online platforms. Yet the theoretically optimal design for estimating many treatment effects remains unclear. We propose a simple Markovian framework for experimental design in which the treatment assignment process is governed by possibly time-varying transition matrices. This formulation encompasses many existing N-of-1 designs and provides a principled way to control temporal dependence in treatment assignment through Markov transition probabilities. Under a finite-order impulse-response model, we formulate the design objective as minimizing the estimation error of ordinary least squares estimators for target treatment effects, and propose practical design optimization procedures. To characterize the optimal temporal structure, we focus on two structured design classes, random-switch and cycle-switch designs, and establish a complete large-$T$ asymptotic theory for the optimal designs in both classes. Our results justify the robustness of i.i.d. Bernoulli designs in N-of-1 trials and quantify how the optimal design depends on the target estimand, including cumulative and lag-specific treatment effects. Simulations demonstrate the effectiveness and robustness of the proposed designs across multiple scenarios.

021

Reposted by Greg Atkinson

John Quackenbush @johnquackenbush.bsky.social · 29/06/2026

Can common analytical methods generate spurious molecular #waves_of_aging? We show LOESS and DE-SWAN produce wave-like patterns from random noise, casting doubt on reported “waves." With MaddyCarbonneau, Kate Shutta, Jeff Miller, Michael Snyder, and Xiaotao Shen. www.biorxiv.org/content/10.6...

biorxiv.org

LOESS and DE-SWAN can induce artifactual "waves" of molecular aging

A growing literature has investigated the relationship between age and biomolecular changes, leading to conclusions that aging occurs in discrete molecular "waves." Data summary tools such as LOESS an...

151

Reposted by Greg Atkinson

Daniel Lakens @lakens.bsky.social · 21/06/2026

New preprint: There is only one correct analysis. osf.io/preprints/ps... with @sajedehra.bsky.social and @mntunc.bsky.social We take a critical look at recent proposals for multiverse or many analyst procedures, and strongly argue that epistemic uncertainty should be reduced, not “embraced”. >

osf.io

OSF

55823

Reposted by Greg Atkinson

University of York Library @uoylibrary.bsky.social · 01/06/2026

Noted, thanks Mel. We should probably hold off on making any firm decisions about REF2029 until we've heard from the bloke who made Angry Birds.

26821

Reposted by Greg Atkinson

Experimental Physiology @expphysiol.bsky.social · 29/05/2026

In this Editorial, Editor-in-Chief Damian Bailey (University of South Wales) discusses restoring Dr Florence Buchanan's missing image in physiology 📖 buff.ly/7u2mBW9

011

Reposted by Greg Atkinson

Nick Tiller, Ph.D. @nbtiller.bsky.social · 28/05/2026

Lock up your wellness grifters. The Health and Wellness Lie is available for preorder. #wellness #misinformation

072

Reposted by Greg Atkinson

Georgia Tomova @georgiatomova.bsky.social · 24/05/2026

Tell me you don't understand Bland-Altman limits of agreement without telling me. Authors: "93.5% (100/107) of pairs fell within 95% limits of agreement"

4135

Reposted by Greg Atkinson

Dan Quintana @dsquintana.bsky.social · 29/04/2026

Selecting an effect size for power analysis is hard. Many researchers fall back on Cohen's thresholds, but they have no empirical basis and vary wildly by field. Our new paper offers a better option: field-specific effect size distributions built from meta-analytic data doi.org/10.3758/s134...

Abstract
Effect sizes are useful for understanding the magnitude of study results and for planning new studies via power analysis.
However, despite their wide usage, effect sizes are often misinterpreted. This is mostly due to an over-reliance on general
effect size benchmarks that were not intended for broad application across diverse research fields. Inaccurate effect size
interpretations can lead to incorrect conclusions about the magnitude of study results and incorrect sample size estimates,
thereby increasing the likelihood of false-positive results. This article introduces the ESDist R package, which is designed
to calculate empirically derived effect-size benchmarks or a range of reliably detectable empirical effect sizes for a specific
research question or field of interest by computing effect size distributions (ESDs). This package can be used on data that
can be easily extracted from pre-existing meta-analyses to help researchers more accurately plan new studies or to better
understand how an individual study might relate to other studies in their field. ESDist includes a set of features that make it
easy to use in a priori power analysis. Moreover, the package includes a feature for estimating effect size benchmarks that
account for publication bias and are weighted by effect sizes' variances, which addresses existing limitations of using ESDs
for study planning or interpretation.

316175

Reposted by Greg Atkinson

Kevin C Klatt, PhD, RD @kcklatt.bsky.social · 16/03/2026

The amount of 'precision' nutrition stuff that is just tossing multi-omics at trial with no real validation, and no real consideration of basic things like inter- vs intra- individual variation is really sad. Just flushing money down the toilet and then putting lipstick on the publication pig

0124

Reposted by Greg Atkinson

Darren Dahly @statsepi.bsky.social · 12/08/2025

A letter from myself and @jdwilko.bsky.social that expands on this a bit more. Thanks to Fertility and Sterility for the opportunity. doi.org/10.1016/j.fe...

Nonrandomized studies of
interventions - complementary or
just convenient?
Schwarze et al. (1) present recommendations to improve the
quality of nonrandomized studies of interventions (NRSIs).
Given the methodological errors we regularly encounter
when reading and reviewing NRSIs in assisted reproductive
technologies (ART), we support the need for such
recommendations to help researchers better design,
analyze, and interpret these studies. That said, we are
concerned that those who require such guidance will,
almost by definition, not understand when the use of
NRSIs is warranted, or mistake them as viable alternatives
to randomized controlled trials (RCTs), despite reassurances
from Schwarze et al. (1) that NRSIs should be viewed as
complementary. Our concerns are particularly relevant for
ART, where regulation is limited, commercial conflicts of
interest are common, and interventions are regularly
adopted before they are evaluated using RCTs. To be clear,
we do not see these as points of disagreement with
Schwarze et al. (1), but rather as additional points for
consideration.

25526

Reposted by Greg Atkinson

Aleksander Molak @alxndrmlk.bsky.social · 02/03/2026

The Optimal Split for an A/B Test Is 50:50 Unless... I recently saw a post explaining why the optimal split between treatment and control groups should be 50:50. The optimal split is indeed 50:50, but only under one assumption: 1/ #CausalSky #StatSky #ABTest #EpiSky #MLSky #EconSky

272

Reposted by Greg Atkinson

Peter Tennant @pwgtennant.bsky.social · 25/02/2026

This recent RCT of an "AI stethoscope" claims the technology "shows promise" for diagnosing cardiovascular conditions. It does not. It is a textbook example of the risks of conducting unprincipled 'per protocol analyses'. Once again, peer review at a major medical journal has failed. 🧵 1/

8436186

Reposted by Greg Atkinson

Experimental Physiology @expphysiol.bsky.social · 04/02/2026

We are delighted to recognise our Exceptional Referees for 2025 🎓 🏆 Visit the webpage below to see the list of referees celebrated for the volume, timeliness, and exceptional quality of their reports & exemplary service to the discipline! 🔗⤵️ buff.ly/H14u9Z1

042

Reposted by Greg Atkinson

Richard D. Morey @richarddmorey.bsky.social · 02/02/2026

New draft: "Decline effects, statistical artifacts, and a meta-analytic paradox". In this manuscript I show how a common practice in meta-analysis (eg the 2015 Open Science Collaboration) creates artifactual signatures of poor scientific behavior. PDF: raw.githubusercontent.com/richarddmore... 1/x

The decline effect (Protzko & Schooler, 2017) is an observed phenomenon where effect sizes in experiments apparently diminish in size from the first paper demonstrating the effect to later replications. This has been taken as a symptom of an unhealthy scientific ecosystem, possibly caused by the "winner's curse" (selection on significance and regression to the mean), publication bias or opportunistic analyses. I show that decline effects can arise as an artifact from a much simpler source: the original article determining the sign of the effect in a meta-analysis. Moreover, such artifactual decline effects will show correlations with some of the same experimental properties that one would expect from biases from poor behavior, such as the sample size of the original study.

78530

Reposted by Greg Atkinson

Experimental Physiology @expphysiol.bsky.social · 14/01/2026

Check out this #EPicks video from James Betts (@uniofbath.bsky.social), as he chats about his article published in @expphysiol.bsky.social which focussed on physiological rhythms and metabolic regulation! 📽️ buff.ly/gpUtCi7 Read the full article here: 📜 buff.ly/rtm1OFd

011

Reposted by Greg Atkinson

Marion Campbell @marionkcampbell.bsky.social · 14/01/2026

Are you passionate about supporting the funding of high-quality methods research in health? If so, please consider applying for the role of Deputy Chair of the NIHR/MRC Better Methods Better Research funding Committee - closing date 22 Jan. See: www.nihr.ac.uk/get-involved...

021

Greg Atkinson @gregatki.bsky.social · 09/01/2026

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...

journals.physiology.org

053

Reposted by Greg Atkinson

Experimental Physiology @expphysiol.bsky.social · 02/01/2026

This #Editorial from Damian Bailey (University of South Wales), David Poole (@kstate.bsky.social) and Ronan Berg (Rigshospitalet) provides an update on the journal from the perspective of our Editor-in-Chief and Deputies for the USA and Europe! 🎓 buff.ly/ogE6Nol

022

Reposted by Greg Atkinson

Triad sou. @triadsou.bsky.social · 30/12/2025

Reflections on the I-squared index for measuring inconsistency in meta-analysis. Julian P. T. Higgins, José A. López-López. Research Synthesis Methods. bit.ly/48Xsjce

bit.ly

Reflections on the I-squared index for measuring inconsistency in meta-analysis | Research Synthesis Methods | Cambridge Core

Reflections on the I-squared index for measuring inconsistency in meta-analysis

042

Reposted by Greg Atkinson

Darren Dahly @statsepi.bsky.social · 29/12/2025

‪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/

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"

1022875

Reposted by Greg Atkinson

Richard Hooper @richardhooper.bsky.social · 19/12/2025

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?

132

Reposted by Greg Atkinson

Stephen Senn @stephensenn.bsky.social · 18/12/2025

Nice piece on the COVID inquiry by Oliver Johnson open.substack.com/pub/bristoli...

open.substack.com

COVID Inquiry (slight return)

The evidence before the court is incontrovertible

152