@georginasturge.bsky.social
Statistics expert and writer, formerly of the House of Commons Library. New book 'Sum of Us' out now ππ Also author of critically acclaimed 'Bad Data' (2022).
I'm thrilled to announce that Talk Data to Me now has a podcast.
My first guest is Jonathan Portes, Professor of Economics at Kings, who talked to me about recent problems with UK economic data & budget challenges for the Chancellor.
Available to listen now π
open.spotify.com/episode/6Txz...
I'm very pleased to share this invitation to this year's Migration Statistics User Forum, which I'll be chairing on 11 December (11:00 to 16:00).
It's a great opportunity to directly question the producers of statistics. And it's online and free. Join us!
www.eventbrite.co.uk/e/migration-...
Nice and clear explainer from the ONS on the changes made to their migration methods today:
blog.ons.gov.uk/2025/11/18/u...
New migration estimates released today reflect major updates to the ONS's methods.
Main takeaways:
* Past British emigration underestimated
* EU emigration overestimated
* New net migration peak of 944,000 in 2022-23, followed by bigger fall than previously thought
www.ons.gov.uk/peoplepopula...
Yeh that claim seems to be about refused asylum refusals specifically, although I haven't yet found published data to back it up. But looking at the Eurostat data on all returns following an order to leave, it was 44/100 last year and a slightly higher 50% over the past three years.
14.11.2025 14:49 β π 1 π 0 π¬ 0 π 0
Could Denmark be a role model for the UK on asylum? This is worth a read if you want to understand the basic challenges each country faces and how well the same approach could copy over.
georginasturge.substack.com/p/fortress-o...
It would certainly be a challenge. The top refused nationalities in Denmark in recent years have been Iranian, Georgian, Syrian, Afghan, and Iraqi. It's only to Georgia that general return rates are anywhere close to 95%.
14.11.2025 12:45 β π 2 π 1 π¬ 0 π 0Eurostat figures show a much lower ratio of all returns relative to orders to return: around 50% over the past three years. I should stress that these are *not* just refused asylum seekers, figures for that specific group are not on Eurostat. But 95% for that group would be quite something.
14.11.2025 11:32 β π 1 π 1 π¬ 1 π 0Also curious to know if anyone has a source for this. I'm seeing it in virtually every news item on Denmark and asylum.
14.11.2025 10:53 β π 5 π 5 π¬ 1 π 0
The 2031 UK census is going ahead after all and the public now has an opportunity to shape how it looks.
What do you think should be asked in the census that isn't in there already? Or are there too many questions as it is?
georginasturge.substack.com/p/if-you-cou...
Milk before water and teabag or after?
12.11.2025 10:16 β π 1 π 0 π¬ 1 π 0
Now that the 2031 UK census is going ahead, the ONS has to decide what's going to be in it.
For that they've asked the public to submit ideas for new questions. Could this be a terrible idea? And what would you like to know about 70 million people?
georginasturge.substack.com/p/if-you-cou...
Now that the 2031 UK census is going ahead, the ONS has to decide what's going to be in it.
For that they've asked the public to submit ideas for new questions. Could this be a terrible idea? And what would you like to know about 70 million people?
georginasturge.substack.com/p/if-you-cou...
The Spectator's Michael Simmons has gathered lots of recent examples of bad data together in this eye-opening piece.
www.spectator.co.uk/article/data...
Why do people *not* respond to the Labour Force Survey? For a while the biggest problem was thought to be failure to make contact with the sampled households.
New data suggests by far the bigger problem - arguably more worrying - is people knowing they've been invited but refusing to take part.
Read more in my latest piece on the faint signs of recovery in UK statistics
georginasturge.substack.com/p/are-there-...
Why do people *not* respond to the Labour Force Survey? For a while the biggest problem was thought to be failure to make contact with the sampled households.
New data suggests by far the bigger problem - arguably more worrying - is people knowing they've been invited but refusing to take part.
Household surveys are vitally important to public policy but response rates are in crisis. Following heavy criticism, the UK's statistical authorities say they're finally getting a grip on the issue.
Might we already be seeing some signs of recovery?
open.substack.com/pub/georgina...
My main feeling about these kinds of poll is that they miss something crucial which is *why* people think we were wrong to leave.
Brexit voters who now say it was the wrong decision are usually signalling discontent with how it was handled, though they'd still back the principle of leaving.
To understand the UK, you must carefully study this map. Tells you all you need to know π₯ βοΈ (I could find a crisps / chips emoji)
15.10.2025 06:26 β π 33 π 5 π¬ 2 π 2
More in my latest substack post, along with reflections on the governmentβs economic data woes.
open.substack.com/pub/georgina...
Thereβs no clear pattern of opposition to digital ID in seats with a high Labour share (almost all of which Labour won). Seems like this has become a great unifier, or perhaps thatβs just people reacting badly to anything announced by Keir Starmer at the momentβ¦?
10.10.2025 12:02 β π 2 π 0 π¬ 1 π 0
I did some quick analysis of who has been signing the e-petition against digital ID and the results are interesting.
Nothing very surprising when comparing signatures per constituency with share voting Reform in 2024.
The picture for Labour is a bit more troublingβ¦
Some thoughts on the Chancellor's data difficulties in my Substack post this week β¬οΈ
Plus, oh yes, I'm still keeping my beady eye on digital ID, and the public reaction so far is not looking good...
open.substack.com/pub/georgina...
Nice view of Somerset House at low tide back when it housed the General Records Office. The embankment now separates the building's facade from the river, and the bridge has been replaced by a bigger concrete one.
09.10.2025 08:03 β π 2 π 0 π¬ 0 π 0
Well this is a new one for me and something that belongs in an update of Bad Data (and in Sum of Us).
Italians mistakenly identified as Roma on Scottish Census.
www.thetimes.com/article/f1ff...
In the sense that it's easier to justify tax increases to plug this bigger deficit than because recent fiscal headroom was riskily small? Might be, but I would think most people probably don't think either is a good reason!
04.10.2025 09:07 β π 1 π 0 π¬ 1 π 0
Economic forecasts are *always* revised & the swing can be substantial. Add to that the OBR choosing this moment to revise productivity figures going back 10 years.
Bad data strikes again, leaving the Chancellor with a fiscal headache & massive PR challenge β¬οΈ
www.thetimes.com/article/83c5...
Very astute piece.
The unification of data systems that was already going on quietly in the background could be at risk now because it's being publicly framed as part of the move to a 'papers please' society.
Timely account of the fascinating history of data in our politics!
It shows how much of the data collection that's defined our sense of who we are as a society across history comes down to a mix of accident, contingency & bias (& plenty of genuine concern for human welfare & a better society too!)
