I had not heard of the recently deceased Dan Simmons but tjere seem to be unusually polarised views of him, both in terms of general quality of his work and racism.

There were HUGE differences between patients who did and did not receive the AI stethoscope. Patients were 20 years older (64 vs 44), 3.5x more likely to have hypertension (53% vs 15%) and 4x more likely to have kidney disease (13% vs 3%) 🧵 11/

imo we could just legislate against this and there is no reason not to

Was the global decline of extreme poverty only due to China? The share of the world population living in extreme poverty has never declined as rapidly as in the past three decades.

A surprising feature of my 'Discover' feed on Bluesky is a bunch of authors in all genres struggling with writer's block, editors, publishers, and various existential crises. I don't believe I have ever expressed interest, and am afraid I have only a distant sense of sympathy, so why am I fed these?

Here, have the thread as a blog post: lizbourke.wordpress.com/2026/02/19/1...

96% of Niueans live outside of Niue (mostly but not all in New Zealand). Cook Islands, Marshall Islands, Tokelau, Samoa and Tonga are also majority diaspora nations. The next #rstats blog post in my series on Pacific population issues is at freerangestats.info/blog/2026/02...

A thing I dislike when part of the audience of a presentation is a final slide that just says "Questions?" or "Thanks" or something similarly contentless. This is the exact moment when it is useful and much better to have a summary slide of the most important or surprising things the speaker said.

This is great example of how to weave text and charts:

Auckland is the world's biggest Pacific Islander city, but which is second? Turns out to be Port Moresby, followed by several Indonesian cities of West New Guinea. Blog post in my series on Pacific people movement issues, with #rstats code, at freerangestats.info/blog/2026/02...

I am genuinely curious about this. Does President Trump now agree with his critics that the bunker busting bombs didn't achieve more than a slight pause to the nuclear program? Does he get asked about this and why more attacks are needed and useful, given the last ones either succeeded or failed?

Using natural language prompting almost certainly makes LLMs a) need to be larger than they could be, b) less efficient and reliable than they could be, and c) much harder to use than they could be.

I covered an R meeting as a science writer a few years ago and it was some of the most fun I've had at a meeting and the attendees were so nice and happy to talk.

I noticed and fixed a mistake in this Pacific net migration plot. The original y axis was labelled as percent when it is in fact per 1,000 resident population. This makes a difference for eg Marshall Islands between losing 50% of its population each year (wrong) and 5%. Still very high, of course!

This is just nuts. In a well-powered randomised trial, giving immunotherapy in the morning rather than afternoon/evening produced ludicrously big improvements in progression-free survival www.nature.com/articles/s41... via www.science.org/content/blog...

Scatter plots are cool! Inspired by @brankomilan.bsky.social's excellent *Visions of Inequality* , I drew some as alternative ways of visualising historic #econsky social tables of income and population by class. Blog post, with #rstats code, at freerangestats.info/blog/2026/02... 1/n

My friend Jon has died | First Dog on the Moon

Actually, #econsky, why is this so? Do any countries have accreditation for economists, and why don't most?

This is probably the single weirdest Biggles book, and is a far cry from fighting the Boche in a Sopwith Camel. It also features a lost tribe who communicate by chirping, people who live for centuries (because of the radium), a swarm of electric centipedes and a sort of magnetic death ray.

Giving the robots Genuine People Personalities was a mistake, and Douglas Adams warned us all.

My hotel room safe is branded as 'Raffles', who does not sound a safe person around one's valuables. But I guess most people won't get the reference, and those who do won't care as far as their safe-use decision-making goes.

Note that this in itself is a well-replicated finding. At least a dozen studies show that when researchers replicate qualitative research, or re-analyze it, they come to basically identical themes. There is nothing special about qualitative research with respect to replicability.

About 3.2 million people live in Puerto Rico as of 2025, down from a peak of 3.8 million in 2004. One reason for Puerto Rico’s declining population is that more people are moving from the island to the mainland U.S. each year than in the other direction.

So much of classic political science research is woefully underpowered. I'm glad we're slowly coming to our senses.

Always love a blog on the late bronze age collapse.

Even putting politics aside, this does sound like one of the worst films of significant budget ever. www.theguardian.com/film/2026/ja...

Devoting tomorrow's lecture to group-level confounding and fixed/random effects. I'm still trying to figure out how I want to present this in next edition of my book. It's just an inherently confusing topic, isn't it? A mix of bad terminology and advanced statistical concepts. I'll keep at it.

#rstats ANNOUNCE: The ASA John M. Chambers Statistical Software Award to Harriet Mason (Monash Univ) For ggdibbler, an R graphics (ggplot2 add-on) package that enables rich visualization of uncertainty across raster, contour, stack, and spatial plots.

Are there any R users working for the Australian federal dept of health? I’d love to hear from you! New gig, new people to meet :)

Sorry I'm a few years late to the party, but I'm here finally! :) yihui.org/en/2026/01/h... Thank you all so much for your generous support over the past two years!

Do predictive models need to be causal? notstatschat.rbind.io/2026/01/26/d...

Basically sound ratings of the Tintin albums. I am more positive on a few but broadly these reviews are very sensible.

"You're on the Island of Conclusions. Make yourself at home. You're apt to be here for some time." "But how did we get here?" asked Milo, who was still a bit puzzled by being there at all. "You jumped, of course," explained Canby. "That's the way most everyone gets here."

The flawed study in question claims corporate sustainability increases stock returns.

Stalin is generally heavily criticised for allowing huge exports including military-relevant material to Germany right up to invasion day, and a common view seems to be he was an idiot for not stopping all trade once it was clear war was inevitable.

Once more about the z-curve method statmodeling.stat.columbia.edu/2026/01/20/o...

In case of interest - the European Commission is open for public submissions for another couple of days (till 22 January) on their 2026 Strategic Foresight Report - "the future of the EU in a changing world". This should be a more interesting edition than most! ec.europa.eu/info/law/bet...

Anyway, here's the post, I'm going to bed now. acoup.blog/2026/01/16/c...

StatsChat on business confidence: in general www.statschat.org.nz/2018/09/02/b... and in particular www.statschat.org.nz/2019/08/07/b...

What’s the essence of blogging? statmodeling.stat.columbia.edu/2026/01/12/w...

Our machine learning competition on forecasting depression is online! We'd love for as many people as possible to participate. Please share in your respective networks — thank you :). www.staff.universiteitleiden.nl/announcement...

After 9 years, I thought it was time for another major release of the forecast package for #rstats! New modelling functions, missing values now handled in ETS models, and a more consistent interface. robjhyndman.com/hyndsight/fo...

Concerns about the z-curve method statmodeling.stat.columbia.edu/2026/01/10/c...

The old view of skilled emigration as a 'brain drain' has collapsed after a generation of research. Skilled migrants cause innovation, technology transfer, and human capital investment in the countries they leave. The best summary is @mushfiq-econ.bsky.social et al.—> doi.org/10.1126/scie...

Excited to share a new R package: 𝗵𝗲𝗮𝘁 🌡️ `heat` makes it easier to work with climate or other gridded data in applied research, providing a comprehensive + optimized set of tools to compute environmental exposures for admin boundaries or points from gridded/point data. github.com/echolab-stan...

🚀 Announcing unicefData — open-source access to 700+ child welfare indicators from UNICEF’s SDMX API One syntax across Python, R, and Stata. No API expertise needed. 🔗 github.com/unicef-drp/unicefData #OpenData #SDGs #RStats #Python #Stata

I blogged with #rstats about New Caledonia's nickel exports and how they have collapsed in the past few years, from a combination of global economic conditions, local political economy, and security and civil unrest problems in 2024. freerangestats.info/blog/2026/01...

LLMs are biased towards their training data. Most published research is...not brilliant. What is the likelihood that an LLM produces a useful hypothesis and not, say, "why not try flax seed oil for liver disease".

For last week's #TidyTuesday, I created a map of endangered languages around the world. Built entirely in R/ggplot2: transforming spatial coordinates, identifying hotspots and composing multiple charts! Code: github.com/gkaramanis/t... Inspiration: www.reddit.com/r/MapPorn/co... #RStats #dataviz