Jack Bailey
@jack-bailey
Political Scientist at the University of Manchester Co-Director of the British Election Study
🇬🇧 UK Poll of Polls: 9 August 2026 27% 🌹 Labour 23% ➡️ Reform 20% 🌳 Conservative 11% 🍃 Green 10% 🐥 Lib Dem 6% 👽 Other 3% ✖️ SNP
If you ever doubt we’re in the banter timeline, remember that “a strong Binface vote” really appeared in a newspaper
There's a language I want to learn that's not on Duolingo, so I made my own app using Codex. I gave it some details about my life to make the expressions more useful, but I feel like it maybe got too close for comfort...
I suspect that @jonathanfreedland.bsky.social is right: the Tories are stuck in the doldrums and won't make it out unless Farage stacks it or they get a new leader. Still, it's also interesting how little impact this fact about the recent GM mayoral election has had.
Greens and Reform come together to block a major solar farm project in Kent 🙃 giftarticle.ft.com/giftarticle/...
Once again we’re in the Arrested Development meme. Did trying to get single market access without free movement of people work in the past? God no … but it might work for us.
Strongly recommend the MC Escher exhibition at Somerset House. Also bonus points to whoever came up with the idea of having the sign for the exhibition at the bottom of the stair well.
That's not right. 2024 was the most fragmented election in UK politics ever. doi.org/10.1080/0140...
Perhaps, but in New Zealand, probably our best comparison, the major parties didn't want PR but the PM misread his briefing notes and committed his party to a referendum in a TV interview. Always important to remember that *politics is weird*! journals.sagepub.com/doi/10.1177/...
🇬🇧 UK Poll of Polls: 2 August 2026 25% 🌹 Labour 23% ➡️ Reform 20% 🌳 Conservative 11% 🍃 Green 11% 🐥 Lib Dem 6% 👽 Other 3% ✖️ SNP
Great to have an opportunity to vote in an election *not* held under first-past-the-post today. Even if it was the supplementary vote 🗳️🐝
I have the following folders in Obsidian. _attachments is just a place where any attachments get stored automatically. _sources contains notes that only include a single Harvard referenced source as the title so that I have something to link to when citing.
When we fret about the effect of new technologies on how we learn, it’s worth remembering that similar concerns have accompanied ~ every new technology *including writing*
🇬🇧 UK Poll of Polls: 26 July 2026 24% ➡️ Reform 23% 🌹 Labour 21% 🌳 Conservative 12% 🍃 Green 11% 🐥 Lib Dem 6% 👽 Other 3% ✖️ SNP
I always remember how affecting I found Jens Stoltenberg's speech after the Utøya massacre where he argued that "the answer to violence is even more democracy". A lesson we should all take to heart.
🇬🇧 UK Poll of Polls: 12 July 2026 24% ➡️ Reform 22% 🌹 Labour 20% 🌳 Conservative 12% 🍃 Green 11% 🐥 Lib Dem 6% 👽 Other 3% ✖️ SNP
🇬🇧 UK Poll of Polls: 5 July 2026 25% ➡️ Reform 21% 🌹 Labour 20% 🌳 Conservative 13% 🍃 Green 12% 🐥 Lib Dem 6% 👽 Other 3% ✖️ SNP
Here it is auditing an average-sized code base. In this case, it's on the money too. Maybe the future is local?!
Empirically, the two measures often move together, but the differences can be meaningful and tend to grow larger where disproportionality is higher. The difference is larger in Australia (high disproportionality), compared to Ireland (medium), or the Netherlands (low). 8/10
With this in mind, I propose a complementary normalised index. Rather than asking “𝘩𝘰𝘸 𝘥𝘪𝘴𝘱𝘳𝘰𝘱𝘰𝘳𝘵𝘪𝘰𝘯𝘢𝘭 𝘸𝘢𝘴 𝘵𝘩𝘪𝘴 𝘦𝘭𝘦𝘤𝘵𝘪𝘰𝘯?”, it asks “𝘩𝘰𝘸 𝘥𝘪𝘴𝘱𝘳𝘰𝘱𝘰𝘳𝘵𝘪𝘰𝘯𝘢𝘭 𝘸𝘢𝘴 𝘵𝘩𝘪𝘴 𝘦𝘭𝘦𝘤𝘵𝘪𝘰𝘯 𝘳𝘦𝘭𝘢𝘵𝘪𝘷𝘦 𝘵𝘰 𝘩𝘰𝘸 𝘥𝘪𝘴𝘱𝘳𝘰𝘱𝘰𝘳𝘵𝘪𝘰𝘯𝘢𝘭 𝘪𝘵 𝘤𝘰𝘶𝘭𝘥 𝘩𝘢𝘷𝘦 𝘣𝘦𝘦𝘯?” 6/10
The key result is that, for an election with N_V effective vote-winning parties, the Gallagher index cannot exceed \sqrt{(N_V + 1)/(2N_V)}. As N_V grows, this upper bound falls towards 1/\sqrt{2}, not 1. 4/10
The Gallagher index measures how far each party’s seat share is from its vote share, then combines the gaps into a score. To max it out, one party must win all votes and another all seats. That's only possible when there is one effective vote-winning party and a different seat-winning party. 2/10
I have a new paper on the Gallagher index: the most common measure of electoral disproportionality. We often assume it runs from 0 (perfect 𝘱𝘳𝘰𝘱𝘰𝘳𝘵𝘪𝘰𝘯𝘢𝘭𝘪𝘵𝘺) to 1 (perfect 𝘥𝘪𝘴𝘱𝘳𝘰𝘱𝘰𝘳𝘵𝘪𝘰𝘯𝘢𝘭𝘪𝘵𝘺). But I show that in most party systems, 1 is not a possible value. journals.sagepub.com/doi/full/10....
The SNP’s brand colour is a pale yellow. The emoji above is a black cross (which is as close to the logo as I can get)
🇬🇧 UK Poll of Polls: 28 June 2026 26% ➡️ Reform 21% 🌹 Labour 19% 🌳 Conservative 12% 🍃 Green 12% 🐥 Lib Dem 6% 👽 Other 3% ✖️ SNP
🇬🇧 UK Poll of Polls: 22 June 2026 26% ➡️ Reform 20% 🌹 Labour 19% 🌳 Conservative 13% 🍃 Green 12% 🐥 Lib Dem 6% 👽 Other 3% ✖️ SNP
