Lightnews — Scholar-powered news

LightNews

Peter Ellis

@freerangestats.info

2.9K followers 1.5K following 2K posts

Director, Statistics for Development Division, at the Pacific Community. Posts are personal views, and most likely to be about data and #rstats, maybe with bits of history, social science and philosophy thrown in. Blog is at https://freerangestats.info/.

freerangestats.info

Posts Media Videos Starter Packs

Pinned

Peter Ellis @freerangestats.info · 21d

Genuine life-changing opportunity to become the Programme Manager for the Pacific Data Hub. Based in lovely Nouméa, New Caledonia; responsible for driving regional data stewardship, innovation and sustainable impact. And free of income tax. careers.spc.int/xMiWgK/progr...

Programme Manager (Pacific Data Hub)

Noumea-based position Attractive expatriate package Join the principal development organisation in the region Description The Pacific Community (SPC) is the ...

careers.spc.int

Reposted by Peter Ellis

Health Nerd @gidmk.bsky.social · 3h

A useful post on the bizarre citation cabal that is the IHME and the Global Burden of Disease study.

Ilya Kashnitsky @ikashnitsky.phd · 3h

I wrapped up my scattered comments on IHME into a blog post

🔗 ikashnitsky.phd/2025/ihme-bibl

tl;dr: avoid getting unwarranted co-authorship recognition and do apply some sort of contribution-weighted thinking when evaluating someone's publication record based on bare numbers

Beyond Fraud: How IHME Distorts Academic Metrics

Dr. Ilya Kashnitsky is a demographer @ Statistics Denmark.

ikashnitsky.phd

Reposted by Peter Ellis

Alexandra de Gendre @adegendre.bsky.social · 11h

A fascinating blog post on the generational loss of reading & its consequences for human knowledge & society —

🔗 open.substack.com/pub/jmarriot...

The dawn of the post-literate society

And the end of civilisation

open.substack.com

Peter Ellis @freerangestats.info · 9h

Yes, a good read. Several things appearing on this recently. I'm not sure if it's right, or if it's important if true because it's so hard to judge this sort of thing as it happens (but I think it's yes and yes). I have been struck recently by how virtually no-one but me seems to read on airplanes.

Reposted by Peter Ellis

Kristen Olson @olsonkm.bsky.social · 10h

Survey costs drive almost every data collection decision, but are woefully understudied. @aapor.bsky.social, @amstatnews.bsky.social, and @westatofficial.bsky.social are cosponsoring this important workshop on Survey Costs in February 2026! Submit your abstract today!

aapor.org/aapor-2026-s...

2026 Survey Costs Workshop - AAPOR

Peter Ellis @freerangestats.info · 1d

A currency swap isn't really 'giving', unless the value of the Argentine currency subsequently drops to zero.

Peter Ellis @freerangestats.info · 1d

Apart from missing lend-lease and such, this person's narrative seems to get the fundamental sequence of even mass troop events wrong. Like, the Germans were still fighting in Stalingrad when operation Torch landed US troops (with others) in the western theatre.

Peter Ellis @freerangestats.info · 1d

This is an easy one. Sick people should stay at home, not sing in choir. Particularly if there's any chance at all it's something infectious.

Peter Ellis @freerangestats.info · 1d

Why did it go up so much in 2019 and 2020?

Peter Ellis @freerangestats.info · 1d

This is a good quick read bringing these things together. It's easy to focus on #3 (prior information). But #4 (regularization), #5 (combining multiple sources of information) and #6 (latent data) are great practical motivations for Bayesian techniques too.

Andrew Gelman et al. @statmodeling.bsky.social · 1d

7 reasons to use Bayesian inference!
statmodeling.stat.columbia.edu/2025/10/11/7...

7 reasons to use Bayesian inference! | Statistical Modeling, Causal Inference, and Social Science

statmodeling.stat.columbia.edu

Peter Ellis @freerangestats.info · 1d

Probably a rabbit hole I'll regret, but what is going on in these guys' heads when they repeatedly talk about >100% discounts? My best guess is "reduce to the point that a 650% increase would be needed to bring it back to current price" ie to 1/7.5 of current. Is this a common kids' mistake?

Aaron Rupar @atrupar.com · 2d

Dr Oz: "Drugs for lung diseases are discounted massively. The president highlighted 650%. Significant discounts ... "

Peter Ellis @freerangestats.info · 3d

Looking at the actual list of winners and thinking that this must have been in 1996 or 1997 given how I knew him, I'm really not sure which group it was. Perhaps it was Amnesty after all.

Peter Ellis @freerangestats.info · 3d

I had a colleague in the 1990s who had "Nobel Peace Prize" listed on top of "Other awards" on their CV, with the explanation "Awarded in XXXX to Amnesty International (or some other similar group, not sure which one it was) and all of its members, including me!".

Reposted by Peter Ellis

econmaett @econmaett.github.io · 3d

Ross Ihaka: The Open-Hearted Data Hero
#rstats
www.historyofdatascience.com/ross-ihaka-t...

“R changed my opinion of humanity to some extent, to see how people are really willing to freely give of themselves and produce something larger than themselves without any thought of personal glory.”

Peter Ellis @freerangestats.info · 3d

This is incredible, great work. I would need multiple passes to produce something as neat and well-structured as this with a computer, never mind by writing by hand like an animal.

Peter Ellis @freerangestats.info · 4d

The only thing that makes this slow is not being in parallel. The "R doesn't work with for loops" has become practically a superstition. Yes, vectorised operations are much better in R than loops, but plenty of operations (like this modelling) don't lend themselves to vectorisation.

Peter Ellis @freerangestats.info · 4d

It becomes a safety issue if person in exit row refuses any instructions, do can't just fix up after.

Peter Ellis @freerangestats.info · 5d

I'm guessing about 100. I got 112 to my work email yesterday and only 6 to personal. But a huge amount of traffic has (rightly) moved to Teams (or Slack, or whatever one uses) these days.

Reposted by Peter Ellis

Guy Berger @econberger.bsky.social · 5d

Bajillionaires of the world, we don't need more employment growth nowcasts during the government shutdown.

But we do need a very-large-sample monthly survey of US households to determine employment / unemployment status.

macromostly.substack.com/p/we-dont-ne...

We Don't Need More Employment Growth Nowcasts!

Slack Nowcasts Are the Missing Piece

macromostly.substack.com

Peter Ellis @freerangestats.info · 5d

This is a more senior role than a similar demography-focused analyst role I posted about a few months ago.

This is a regional leadership role, with significant stakeholder engagement and capability building role in addition to the technical work.

Peter Ellis @freerangestats.info · 5d

We're #hiring a demographer at Pacific Community-SPC to help Pacific Island countries and territories with their population statistics! Based in Fiji, we're seeking someone with a relevant Masters or PhD, 10+ years experience, and strong skills in #rstats or Stata. careers.spc.int/job/statisti...

Statistics Adviser (Demography)

Suva-based position (Fiji) Attractive expatriate package Join the principal development organisation in the region Description The Pacific Community (SPC) is...

careers.spc.int

Peter Ellis @freerangestats.info · 5d

'No kings' is a terrible slogan for this as it distracts from the actual issue. Plenty of kings are well bound by constitution, law, and process, and plenty of presidents and dictators aren't.

Peter Ellis @freerangestats.info · 5d

To be fair to Cruz, I think these things genuinely are different. Politicians as legislators are obviously free to engage in public debate, including asking influential organisations to explain themselves. It's completely different when the executive does it.

Reposted by Peter Ellis

Jenn Dowd @drjenndowd.bsky.social · 6d

2/ but I think social media is a good place to practice clear & concise communication of work you think is important (your own & others). And of course it has the chance to amplify the reach & understanding of that science much more than the readership of academic journals.

Peter Ellis @freerangestats.info · 6d

I guess we should be grateful there's no sequels to E.T. or Close Encounters of the Third Kind.

Peter Ellis @freerangestats.info · 6d

Yes definitely, good addition.