That doesn't make RCTs "bad" but we could probably rethink how we interpret them with standard estimands. The ATE may not be all that important!

PM resignation watch followed by a heat health alert is peak UK news cycle

One thing I think the focus on block politics underplays are the (small) voters who 'least liked' party are in their 'block'.

Cannot recommend this documentary enough ! www.bbc.co.uk/iplayer/epis...

This was the 2024 result in Makerfield, where Josh Simons is standing down to allow Andy Burnham to return to Parliament. It's going to be really, really tough for Burnham to win it.

This whole by-election will now be about the future PM. Have we ever seen a by-election like it?! It will change the calculations, divert so many resources there, make the choice high stakes. Maybe even a slam dunk result that shows Reform's vulnerabilities. A killer move, if so, for Burnham.

One of the largest drops in nhs waiting lists on record, which no one will now hear about. Opportunity cost from all the leadership turmoil www.england.nhs.uk/2026/05/heal...

The local elections in one chart. Labour has been squeezed between the Greens in young wards and Reform in older working-class wards www.economist.com/britain/2026...

Early individual-level survey data from around the local elections 👇 allows us to start understanding what drove the huge LAB (& CON) losses, which the ward-level data we've seen so far just can't tell us

Nationally, this suggests Labour vote disproportionately stayed at home or went to another left party. Massive collapse to Plaid in Wales. Reform turned out an energised base with some switchers on top. Greens also. Lab->Reform exists but not the major part of the picture.

Does anyone know of (/have) individual-level data on "split ballot" voting (i.e. voting for candidates from different parties) in English local elections?

Campaign Lab are hiring for 3 new roles - Great opportunity for anyone from an academic or applied research background who's excited about running RCTs and doing data science with progressive campaigns in the UK! campaignlab.uk/work-with-us/

🧵/ Let's put Labour's local election results into perspective... They lost 58% of the seats they were defending, the 10th worst performance for any party in the last sixty years and the 5th worst for a govt. This compares to the Tories losing 48-53% in the locals prior to their landslide defeats.

👇

Overcoming the ecological fallacy by appealing to the world spirit

This crossed my mind a lot when looking at the locals in 2025, where the raw votes in many wards suggested either significant switching from left to right, or Reform gaining non-voters and Labour voters staying home. The individual-level data then showed that intra-bloc switching was low... 1/2

Ahead of the local elections, here's an easy guide of how best to assess the results as they come in, including the best metrics to use and some historical benchmarks. It's on the longer side, but is split into sections and filled with lots of (hopefully) useful data, and completely free-to-read.

British politics is resembling the rest of Europe: • Multi-party competition has replaced two-party dominance • Challengers are on the rise, esp. the populist right • Political realignment centres on identity politics (incl Brexit) But FPTP is poorly suited to this fragmentation

This but "ecological fallacy"

Really helpful framework for thinking about the utility of survey experiments for practitioners by @benmtappin.bsky.social 👇 www.benmtappin.com/publication/...

🧵1/ Our first meta-science paper (with 350+ coauthors) is published today in Nature. It presents one of the largest-ever reproducibility projects in economics & political science. Here’s what we found 👇

Not interested 😂

2-3 on aggregate 4-1 on the posts #CHEARS #UWCL

Delighted our paper (with an amazing group of co-authors) is forthcoming in the APSR Takeaway for Labour and other centre-left parties: fixing public services is key to reducing support for the populist right 👇

📄Published Today in Nature: 500 researchers reproduced 100 studies across the social & behavioral sciences to assess their analytical robustness (led by @balazsaczel.bsky.social & @szaszibarnabas.bsky.social). Article: www.nature.com/articles/s41... Preprint: osf.io/preprints/me... TLDR: 1/11

🚨 New POAL brief from @victoraraujo.bsky.social & @miriamsorace.bsky.social out now: how to properly evaluate the causal impact of shocks, debates, and political crises on public opinion. Diagnostics, pitfalls, examples … all in three pages! Read here: www.poal.co.uk/POAL_brief_U...

AJPS article from @kenbenoit.bsky.social, Scott de Marchi, Conor Laver, Michael Laver, and Jinshuai Ma on using LLMs to analyse political texts. 🧵 1/8 Social science research requires reliable extraction of information from texts to identify authors' preferences on various policies and issues.

Turns out sorting below the threshold is quite hard

HOPE not hate stands in solidarity with Jewish communities in Golders Green and around the country after last night's arson attack.

This cool website tracks and collates union wins & achievements (partially) through automated means - fun to browse & could make for a really interesting dataset! watcha.site

The call for papers for #EPOP2026 in Liverpool (17–19 September) is now open! We welcome submissions across all areas of research on elections, public opinion and parties. Deadline: 26 April Submit your proposals here: www.liverpool.ac.uk/politics/eve... @polstudiesassoc.bsky.social

As @bwalker.uk points out at the other place, progressives being undecided until late *could* be coming up in polls as they will be Don't Knows and will need squeezing for vote intention. We already know that DKs look quite a bit like Labour voters in policy views. So an undercount on the left?

Important data about Gorton & Denton from @bwalker.uk. 1. Only around 5% of Lab 2024 voters switched to Reform vs 36% to Greens. 2. No evidence of support for Reform among under 35s (sub 10%) - it's all over 55s 3. Progressives were undecided til last minute www.newstatesman.com/politics/pol...

In 2024, 744 parties competed in the Indian general election, 10 times as many as in the UK. Yet we describe both as 2-party systems. Why? My new post explains how to count like a political scientist: using the *effective* number of parties. jack-bailey.co.uk/blog/what-is...

A headline in the Times was inaccurate and they have issued this correction: “YouGov is not making changes to its methodology or to any previously published results. We are happy to make this clear” We are not changing our methodology or results, which accurately predicted Reform's vote at GE2024

If I had to guess (always a bad idea) I suspect there would be changes to the freedom of expression and freedom of association components following on from the Public Order Act 2023 and use of the proscription powers under the Terrorism Act (2/2)

The UK decline (shown below) is particularly marked. It's surprising to see a decline to below levels of the 1970s, when there was (for example) internment without trial in parts of the UK (1/2)

A bit of data science, a bit of social history, what more could you ask for? Another excellent blog by @laurenleek.eu open.substack.com/pub/laurenle...

Oh no, someone left the testing constituency in!

New w/@scottclifford.bsky.social. Lots of work uses agree-disagree scales, and a lit review shows these are 1) frequently just measured in one direction (agree = higher trait) and 2) correlated with each other. This has potentially big issues for conclusions. link.springer.com/article/10.1...

This is really worth reading, not least as it highlights how the 'Hero Voter' idea has been twisted beyond all recognition in most commentary and Westminster briefing.

As Maria Sobolewska and I first said in Brexitland, the legacy parties under FPP are like Tinkerbell - they need belief to survive. if people cease to believe they see the best and only options, they can die fast. Is this the moment Labour’s Tinkerbell dies?

Well, no, they’re not just as wrong. The bigger error is not seeing that “trying to win over voters from a party on the opposite end of the spectrum, whose values your core voters oppose” is not same as “chasing voters from party on same side ideologically whose values your core voters like.”

Here's a table in the new @britishelectionstudy.com book (forthcoming). We calculated 'second preferences' - here Labour voters after elections 2015 to 2024. 50% of Labour's 2024 voters had Greens as second preference. 42% the Lib Dems. The left bloc is coalescing behind the most viable left party!

Labour should have focused the campaign more on what voters really care about: the licensing terms of critical national geospatial datasets.

Maybe Matt Goodwin is just conducting the mother of all ethnographies