I am on paternity leave, so not promoting this properly, but lots of Britons have stable but idiosyncratic views , as shown in this book doi.org/10.1093/9780...

Yes that’s definitely true. I have a lot of cites to my “only use this method if you run a validation for your use case” papers that read “Mellon (2014) shows that this method is valid for my use case”

I think the LLM hallucinated citations problem is going away (because it’s fundamentally easy to check if a paper exists using an API) but the much harder problem (especially for faculty reviewing student work) will be hallucinating the content of citations

It's a bit annoying that the default python setup for datascience (jupyter notebooks) is a bad fit for git version control (you end up with inconsequential chunk output merge issues). I wish they'd gone with a stricter separation of code and output like rmd.

If it makes you feel better seahorses are more closely related to horses than sharks

Psychologists stop trying to do psychological science no wait

When you stop borrowing our methodologies

Economists stop trying to do political science

Tim Harford has a great podcast on this where he shows that people didn’t lose that much money on the tulip bubble but lost their shirts investing in railroads (despite the fact that those were a transformative innovation) timharford.com/2023/01/caut...

By the way this is separate from whether you think AI companies stocks are overvalued. The technology at its present level is going to reshape academic research even if Sam Altman goes bankrupt

I know Bluesky is more LLM sceptical, but I definitely had the experience both that gpt-5.2 is superhuman at getting proofs to compile and that it’s not that well aligned (and is chasing compilation success not proving the substantive claims I wanted to)

I've been playing with Lean to prove consistency of a 2-stage estimator in a sociology paper. I have a question for people who do lean proofs. If you understand the theorem and the assumptions it relies on, how much can you treat the an LLM lean proof as a black box

I’ve had an idea floating around my head that there might be a niche audience that would be interested in watching me live stream researching and writing social science papers in realtime on twitch. Terrible idea or kind of fun?

I'm sure someone else has suggested this, but it's time for mandatory DOIs in reference lists. This will make it easier to identify hallucinated references

Does anyone see an issue with using parenthesis color highlighting when typesetting equations like this in latex? ChatGPT seems to think it’s uncouth but I think it would be a clear improvement.

Yeah basically agree with that. Memorising syntax, I'm relatively sure, will become like punch cards

Yes but someone has to also evaluate the foundations of some set of competing approaches. Maybe it's fine and we can just trust LLMs to do that faultlessly. But I fear we're in an odd limbo land rn. Even SOTA models can still do some mad things

Expanding on this. The rise of LLMs push the role of data scientists closer to systems engineers and away from coders. Systems engineering is about turning competing stakeholder demands into consensus criteria for judging a solution and coming up with ways to verify whether criteria have been met.

The industry data science and research jobs that academics have been (maybe) preparing students for over the last decade or so will be unrecognizable in the next 3 to 6 months because of AI. It's coming that quickly and the level of change in day-to-day work is unprecedented.

Gemini3-pro has a weirdly condescending personality.

Does anyone else feel like there’s a lot more respiratory diseases going around this winter? Is there any data backing this up or is everyone I know unusually unlucky?

I’ve been running a history book club for the last few months. What are everyone’s favorite history books?

It’s a real job! But sadly it actually involves things making classifications of diseases and medicines

One day I want to create a wrapper package for lm called uniform_prior_lm that does nothing but relabel CIs and p-values as their Bayesian equivalents.

Camus clearly never had to work out what was wrong with district level outcomes at the 2007 Burkina Faso legislative elections

New blog post! Let's say you have two measures meant to capture the same confounder. They're highly correlated. Can you still proceed with your regression analysis? (I admit, the title is a bit of a spoiler) www.the100.ci/2025/10/13/i...

Has anyone had success stories (or can point me to writeups) with fine tuning frontier LLMs?

Education? Completed it, mate. Last week, I defended my PhD thesis and am now a doctor! Thank you to @robfordmancs.bsky.social and @jonmellon.bsky.social for being great supervisors for 4 years And thanks to @sarahobolt.bsky.social and Ed Fieldhouse for their interesting feedback and discussion

Rex W. Douglas PhD Applied Scientist (Remote/Austin) Looking for full time and freelance projects. Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.

Help! ─ 𝗔𝗽𝗽𝗹𝗶𝗲𝗱 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁 open to work ─ Remote/Austin → www.rexdouglass.com With new tools, I've been cooking: ▸ Machine-vision pipelines — RIOS Intelligent Machines ▸ Information-extraction pipelines — Microsoft ▸ Interactive SMS surveys — Pantheon Insights

I remember around 1995 a woman knocked on the door, doing a survey about air fresheners. I never bought any so it didn't last long but I was quite excited that someone was asking my opinion about a product or service. At last! Finally, Gen X was being consulted!

Social scientists often bemoan falling survey response rates. A big part of the story has to be the proliferation of survey requests. I took every survey I was invited to in an 8 day period and took 27 surveys totaling 33 minutes.

I would argue this is actually about Bayesian inference

Come see our work on using LLMs to track political science research methods at APSA at 4pm today! We find the latest LLMs beat graduate student RAs at extracting information from academic papers

🚨DATA RELEASE 🚨 The BES team are pleased to announce the release of Wave 30 of the British Election Study Internet Panel. Please follow the link below, and we look forward to seeing your research! www.britishelectionstudy.com/bes-resource...

Now elsevier have been harassing me personally to pay the invoice (several thousand dollars).

Has anyone else had experience with being harassed by @elsevierconnect.bsky.social over open access payments?

The lovely people @britishelectionstudy.com released the random probaility data which means we can begin to unravel some unanswered questions about turnout in 2024. A substack to follow but a starter of social class and vote/non-vote. Which party represents the working class? None of the above.

5. Today I read a paper by @sabinaleonelli.bsky.social and Alexander Mussgnug that I think illustrates this point perfectly. philsci-archive.pitt.edu/24891/1/Phil...

Does anyone have a good reference for the theory that the polls overestimated remain in 2016 because of a lack of education weighting? I've heard this theory but not sure if anyone ever fleshed it out in writing

can anyone see a downside to the plan of ordering this desktop specifically for the narrow set of workflows where I have really high RAM requirements and using a modern Macbook pro for every thing else?

Has anyone written a paper on the ideology/partisanship of pollsters?

I’ve never really understood why they don’t force separate tokens for digits. Unlike words, the individual characters of numbers are always meaningful in a consistent way

Yes! I recommend validating LLMs against a task you're actually interested in rather than trying to guess whether these failure modes indicate that it will or won't be capable.

The one grammar rule I refuse to follow is putting a full stop within a quotation rather than after it. It just feels like misquoting the original source (it changes the meaning of the original by implying the sentence ended there).

PSA: Just trotting out the aphorism that “all models are wrong, but some are useful” doesn’t mean *your* model is among the latter…

If researchers find Cohen’s d = 8, no they didn’t. (probably.) Thoughts on the distribution of Cohen's d values in psychology and why we should think about their plausibility when reading articles.

Extreme heat during the growing season increases the number of undernourished households in the following year, measured by calories, iron, and other nutrients in rural India, from Paul Stainier, Manisha Shah, and Alan Barreca https://www.nber.org/papers/w34047

In our polarized political environment, I’m excited to share my new ASR publication: “Competence over Partisanship: Party Affiliation Does Not Affect the Selection of School District Superintendents”. journals.sagepub.com/doi/10.1177/...