We had a good run

A lot of talk about Overton windows after one of the worst years for out-and-out racism & hate in politics that I personally remember - but accusations of racism by Rfm candidates last year was met with withdrawal and blaming "bad vetting". They are far more willing to back racist candidates now.

🚨 Excited that "Making Bribery Profitable Again? The Market Effects of Suspending Accountability for Overseas Bribery" by Eddy Malesky, Lucio Picci, and me has been published in FirstView at @iojournal.bsky.social! Open Access: doi.org/10.1017/S002... 1/n

In case anyone else is doing #adventofcode in R this year, I've been putting my solutions in this repo: github.com/alan-turing-...

🥹

Holy shit. Reuters reporting that new admin instructions on visas are if you worked at a platform in trust & safety or content moderation or on fact checking or online safety at an platform you *and your loved ones* are ineligible for H-1B visa. www.reuters.com/world/us/tru...

Some professional news: I've been promoted to senior data scientist at the Alan Turing Institute

No real policy motive from No 10 for bringing numbers down further, but is there a political one? Our @ukandeu.bsky.social-funded survey research on net migration says no - small benefit of getting numbers from peak to ~350k but none from further cuts bsky.app/profile/robj...

Anecdotally I've found that GitHub Copilot can be an excellent tool for: 1) specific, narrow tasks 2) documentation (esp. autocomplete on e.g. Python docstrings) 3) debugging Also found ChatGPT v. useful for quickly grokking the APIs of under-documented libraries

I'd add a fourth - politicians/elites have an inaccurate picture of public sentiment on immigration. People will write whole articles about how we Must Do X to See Off Farage - and then when you ask what evidence there is that that would achieve that, just silence.

There are three basic problems with our immigration discourse. First, it is fundamentally dehumanising. Second, a la @sundersays.bsky.social, it has left the public with an inaccurate picture of immigration. Third, it treats immigration in isolation, rather than as connected to other policy areas

Again some excellent work by @philswatton.bsky.social!! Check it out:

This whole thread is a wonderful defence of a humane and tolerant attitude, but how prescient was Stuart Hall here: " The capacity to live with difference is ... the coming question of the twenty-first century."

If, like me, you are an enjoyer of Ben's patented bubble plots, you'll have likely noticed that voters cluster by ideology and not demographics. My new preprint is on exactly that: osf.io/preprints/so...

This is every bit as good as people have said. Well worth a read

Headline empirical findings: Models using demographics are not much better at prediction than just assuming all voters vote the same way. Models using ideology however are better, and are close to models using past vote choice w/ constituency vars This result generalises from BES to EES

I have a new preprint: 'Predictive Modelling Shows Demographics Do Not Predict Vote Choice at the Individual Level' DOI: doi.org/10.31235/osf... Thread below

I have a new preprint: 'Predictive Modelling Shows Demographics Do Not Predict Vote Choice at the Individual Level' DOI: doi.org/10.31235/osf... Thread below

It's not even the case that Down's model suggested the median voter would be vote-optimising. It was only such if the single ideological dimension of public preferences is normally distributed.

Morally repugnant.

Reposting this for I suspect not the last time in the lifetime of this parliament: renewal.org.uk/blog/public-...

Outsourcing an ML results interpretation problem: You are comparing classifiers A and B in terms of accuracy on a small number of datasets (say, 15) with small test sets in the range of 100-200 obs. (1/3)

I've written a blog post about using metric unfolding to extract latent dimensions from data, using the ANES feeling thermometers & BES 'like' scales as examples. It's a high-level intro to unfolding and using the smacof R package. These plots are the results. philswatton.github.io/2025/11/05/e...

"If the person is a U.S. citizen or otherwise lawfully in the United States, that individual will be free to go after the brief encounter." --Brett Kavanaugh, Noem v. Vasquez Perdomo www.chicagotribune.com/2025/11/03/c...

Great piece, very much looking forward to part 2. Excellent critiques both of various right-wing ideologies around tech and of left-wing disinterest in tech.

The UK's liberal democracy is not being defended by its mainstream politicians in Labour or the Conservatives. Only third parties seem alive to the threat. Unless something changes, and soon, we will fibd ourselves experiencing our own version of Trump's US.

🚨NEW BLOGPOST🚨 Since the government can't seem to see it, I explain how and why a simple act of Parliament can bring the Andrew affair to the end - and why they should do it. joxleywrites.substack.com/p/parliament...

One of my issues with Politico is that its jaunty, gossipy tone jars with the dangerous turn in UK politics. A "mostly clanger free conference"? A recording of the Shadow Justice Secretary complaining about seeing "no white faces" surfaced and was the talking point in interviews and coverage...

Just caught up on this fantastic essay by @sundersays.bsky.social in @renewaljournal.bsky.social. Brilliant, punchy, to the point on Labour's lack of serious engagement with multiculturalism, and a much-welcome rebuttal to those who advocate for Labour to accommodate Reform's radical right populism.

I wrote a slightly longer response to the NS piece on my substack. You can read it here: dysfunctionalprogramming.substack.com/p/some-comme...

The thing I most hate about this kind of reasoning is that it regards the question of whether or not Farage is a racist as unimportant or incidental. All that matters is what an imagined median voter likes right now. Politics as anything other than triangulation regarded as a misstep.

Short article covering the psychometrics and LLMs work we published a few months back. Hopefully, more readable (i.e. less maths!) :) @philswatton.bsky.social cetas.turing.ac.uk/publications...

Quite. So many Labour people on this website visibly have no idea of just how bad things have got for ethnic minorities since they came on, and how their visible indifference lands.

Brilliantly lucid piece from Phil - well worth a read.

New piece by me. I wrote about what the political methodology literature tells us about the uses and abuses of surveys, why a polling-first political strategy is a mistake, and why a clear set of values and a clear political vision is essential for reconciling contradictory pressures on policy.

It's good to see these MPs continuing to stand up for the role of deliberation in democracy and for the scrutinising role of parliament. If Farage does become PM in the future, a concentration of power in the executive will prove a catastrophe.

really feels like housing affordability is one of the most important political issues of our time, affecting nearly every aspect of our lives, and everyday our mainstream political discourse is about whether sour cream is woke or not

A pandas quirk I recently discovered:

An *amazing* opportunity to work at @gesis.org as a senior researcher with the #CSES project and have a big impact in the field of voting behavior and elections!! Deadline for sending in applications is August 15! #PoliSciJobs

New YouGov polling. A monumental failure of our political class to educate, a monumental failure of our media to report fairly, for a generation

Sarah Mercer, Daniel P. Martin, Phil Swatton Applying Psychometrics to Large Language Model Simulated Populations: Recreating the HEXACO Personality Inventory Experiment with Generative Agents https://arxiv.org/abs/2508.00742

Had a lot of fun collaborating on this project, exploring what personality dimensions are simulated by a population of LLM agents. A big thank you to Sarah for involving me

Paid research assistant opportunity with the British Election Study for UK-based graduate students. Help us validate turnout in our surveys against Electoral Commission records

Wow—the Lemkin Institute for Genocide Prevention (named after Raphael Lemkin, a Holocaust survivor who coined the term genocide) says plainly: "No more discussions, meetings, reports, sessions, summits…IF YOU HAVE THE POWER, YOU MUST USE IT." "HISTORY WILL DETERMINE THE RECTITUDE OF YOUR ACTIONS."

The Pope has a math degree, it should be noted. Probably the first Pope who knows what a metric space is.

I've written about how the party went from tightly controlling selections to tightly controlling the whip, and the limits of that strategy when you a) win huge numbers of seats b) do things your base doesn't like and c) have famous MPs like Abbott. www.newstatesman.com/uncategorize...

I'm trying to think about how I would have reacted if shown this headline five years ago

I do wonder about the longevity of the business model of OpenAI et al. when others are open-sourcing LLMs: ethz.ch/en/news-and-...

📣 NEW PAPER ALERT! 🚨 "School subject choices in adolescence affect political party support" Just published in @wepsocial.bsky.social with @nspmartin.bsky.social and @rolandkappe.bsky.social. doi.org/10.1080/0140... 🧵👇