James Hole
@jameshole02.bsky.social
Maths grad doing data science in aviation.
Interested in govt, public policy, polling, epidemiology, sport, music, and coffee!
Interested in govt, public policy, polling, epidemiology, sport, music, and coffee!
Reposted by James Hole
Nothing but respect if Reeves says "shall we all just have the hour back if everyone's read it?"
November 26, 2025 at 12:20 PM
Nothing but respect if Reeves says "shall we all just have the hour back if everyone's read it?"
Reposted by James Hole
It's Friday afternoon, so who wouldn't want to enter the weekend with one more round of my patented bubble plots of British politics? In this final (one hopes) of my mini-series I look at who the 'odd ones out' of British politics are. Who is 'out of touch' with the rest of the public. Let's see 1/n
November 21, 2025 at 4:15 PM
It's Friday afternoon, so who wouldn't want to enter the weekend with one more round of my patented bubble plots of British politics? In this final (one hopes) of my mini-series I look at who the 'odd ones out' of British politics are. Who is 'out of touch' with the rest of the public. Let's see 1/n
Reposted by James Hole
Can we go back to the days when politicians said that they really wanted to help people but it’s very hard and takes time, rather than saying that they want to hurt people and can do it quickly?
November 16, 2025 at 5:53 PM
Can we go back to the days when politicians said that they really wanted to help people but it’s very hard and takes time, rather than saying that they want to hurt people and can do it quickly?
Reposted by James Hole
Just because an LLM can produce a report with various figures & charts doesn't mean it is good at statistics.
Because good statistics is not about producing code.
It's about deep knowledge of study design & conduct. In my opinion, 95% of all data science problems come from poor questions & design.
Because good statistics is not about producing code.
It's about deep knowledge of study design & conduct. In my opinion, 95% of all data science problems come from poor questions & design.
Y'all. I just got ChatGPT to do everything in R for this manuscript. I mean EVERYTHING. And it's all legit and reproducible. I'm shook.
How are we mentoring our trainees in statistics now? Who needs to learn coding in R line by line, and who doesn't?
scienceforeveryone.science/statistics-i...
How are we mentoring our trainees in statistics now? Who needs to learn coding in R line by line, and who doesn't?
scienceforeveryone.science/statistics-i...
Statistics in the era of AI
How do we mentor, teach, and do stats when AI can do so much of the work?
scienceforeveryone.science
October 9, 2025 at 1:17 PM
Just because an LLM can produce a report with various figures & charts doesn't mean it is good at statistics.
Because good statistics is not about producing code.
It's about deep knowledge of study design & conduct. In my opinion, 95% of all data science problems come from poor questions & design.
Because good statistics is not about producing code.
It's about deep knowledge of study design & conduct. In my opinion, 95% of all data science problems come from poor questions & design.
Reposted by James Hole
They should really just publish the spending review documents.
Not for me to actually read of course, god no, I just want others to crack on with their articles and substacks.
Not for me to actually read of course, god no, I just want others to crack on with their articles and substacks.
June 11, 2025 at 12:04 PM
They should really just publish the spending review documents.
Not for me to actually read of course, god no, I just want others to crack on with their articles and substacks.
Not for me to actually read of course, god no, I just want others to crack on with their articles and substacks.
Reposted by James Hole
Within the last five years the government was capable of making it illegal to have your girlfriend stay over for a night.
This talk strikes me as being a cover for (a) a lack of ideas about how to wield power (b) poor ministers, and (c) the very existence of trade-offs.
This talk strikes me as being a cover for (a) a lack of ideas about how to wield power (b) poor ministers, and (c) the very existence of trade-offs.
Simon Case, recent cabinet secretary, to Harry Cole re OBR and giving power away
June 11, 2025 at 8:10 AM
Within the last five years the government was capable of making it illegal to have your girlfriend stay over for a night.
This talk strikes me as being a cover for (a) a lack of ideas about how to wield power (b) poor ministers, and (c) the very existence of trade-offs.
This talk strikes me as being a cover for (a) a lack of ideas about how to wield power (b) poor ministers, and (c) the very existence of trade-offs.
Reposted by James Hole
Iconic tap tackle by Luke Cowan-Dickie in the rugby just now
March 15, 2025 at 5:16 PM
Iconic tap tackle by Luke Cowan-Dickie in the rugby just now
Reposted by James Hole
Real "the doge ate my homework" energy
February 15, 2025 at 10:51 AM
Real "the doge ate my homework" energy
Reposted by James Hole
If you want to win an election against far right opposition you need to bring together two groups. The activated liberals outraged by the other side and the more moderate groups who want govt to make their lives less difficult. Not by trying to win the enthusiastic supporters of the other side.
February 9, 2025 at 10:22 AM
If you want to win an election against far right opposition you need to bring together two groups. The activated liberals outraged by the other side and the more moderate groups who want govt to make their lives less difficult. Not by trying to win the enthusiastic supporters of the other side.
Reposted by James Hole
This rugby match is a great answer to the question ‘what would it look like if people played some nice free-flowing rugby while having forgotten how to catch?’
February 8, 2025 at 5:19 PM
This rugby match is a great answer to the question ‘what would it look like if people played some nice free-flowing rugby while having forgotten how to catch?’
Reposted by James Hole
New post just out:
"Can Labour Fix the NHS?"
They've done it before but now there's less money and everything's in more of a mess. So it's going to require a different approach.
(£/free trial)
open.substack.com/pub/samf/p/c...
"Can Labour Fix the NHS?"
They've done it before but now there's less money and everything's in more of a mess. So it's going to require a different approach.
(£/free trial)
open.substack.com/pub/samf/p/c...
Can Labour Fix the NHS?
Labour strategists believe two things have to happen if they’re to win a second term.
open.substack.com
February 2, 2025 at 9:03 AM
New post just out:
"Can Labour Fix the NHS?"
They've done it before but now there's less money and everything's in more of a mess. So it's going to require a different approach.
(£/free trial)
open.substack.com/pub/samf/p/c...
"Can Labour Fix the NHS?"
They've done it before but now there's less money and everything's in more of a mess. So it's going to require a different approach.
(£/free trial)
open.substack.com/pub/samf/p/c...
Reposted by James Hole
Great, intense, first half in Dublin. Love the six nations.
February 1, 2025 at 5:34 PM
Great, intense, first half in Dublin. Love the six nations.
Reposted by James Hole
On Thursday, the govt will publish its Plan for Change. This will include key 'milestones' for each of the govt's five missions. Targets such as these have been used extensively by previous govts trying to improve public service performance. But how effective are targets?
A short🧵on the evidence:
A short🧵on the evidence:
December 2, 2024 at 8:37 PM
On Thursday, the govt will publish its Plan for Change. This will include key 'milestones' for each of the govt's five missions. Targets such as these have been used extensively by previous govts trying to improve public service performance. But how effective are targets?
A short🧵on the evidence:
A short🧵on the evidence:
Reposted by James Hole
This is the important point. Yes all polls within the 'margin of error' but that should be a random distribution of error not everyone wrong in the same way.
With a few notable exceptions (incl. Selzer) most polls in the US were fairly close to the election result. What's concerning is that almost all polls across the country understated Trump's advantage. The bias is not huge, but it is systematic.
November 8, 2024 at 11:06 AM
This is the important point. Yes all polls within the 'margin of error' but that should be a random distribution of error not everyone wrong in the same way.
Reposted by James Hole
I have a joke about causality, but I'm not sure what the effect will be.
I have a joke about experimental social psychology, but it doesn’t work if you repeat it.
I have a joke about scientific publishing, but I need to run it past my peers first.
October 27, 2024 at 9:20 AM
I have a joke about causality, but I'm not sure what the effect will be.
Reposted by James Hole
Something a bit different to start the week: Looking at the 'two-party' share 2015 - 2024 and Labour's share of that two-party share. 2024 isn't a two party story but adding the context since 2015 shows how it changed from fragmented to two-party and back again.
October 14, 2024 at 10:01 AM
Something a bit different to start the week: Looking at the 'two-party' share 2015 - 2024 and Labour's share of that two-party share. 2024 isn't a two party story but adding the context since 2015 shows how it changed from fragmented to two-party and back again.
Reposted by James Hole
England's criminal courts are in disarray. If you are victim of a crime now, and the police catch the suspect and the CPS charges them, the trial may not take place until 2027.
My essay for this weekend's FT on how this happened and what might fix it on.ft.com/401MjpS
My essay for this weekend's FT on how this happened and what might fix it on.ft.com/401MjpS
How the English courts reached breaking point
[FREE TO READ] A record backlog of criminal trials has left lawyers ‘drowning in cases’. Henry Mance goes in search of the answers
on.ft.com
October 13, 2024 at 9:58 AM
England's criminal courts are in disarray. If you are victim of a crime now, and the police catch the suspect and the CPS charges them, the trial may not take place until 2027.
My essay for this weekend's FT on how this happened and what might fix it on.ft.com/401MjpS
My essay for this weekend's FT on how this happened and what might fix it on.ft.com/401MjpS
Reposted by James Hole
New post just out:
"The Efficiency Paradox"
My explanation for why governments find it so hard to save money and how they could make big savings in the future.
(£/free trial)
open.substack.com/pub/samf/p/t...
"The Efficiency Paradox"
My explanation for why governments find it so hard to save money and how they could make big savings in the future.
(£/free trial)
open.substack.com/pub/samf/p/t...
The Efficiency Paradox
Why are governments so bad at saving money?
open.substack.com
September 19, 2024 at 7:39 AM
New post just out:
"The Efficiency Paradox"
My explanation for why governments find it so hard to save money and how they could make big savings in the future.
(£/free trial)
open.substack.com/pub/samf/p/t...
"The Efficiency Paradox"
My explanation for why governments find it so hard to save money and how they could make big savings in the future.
(£/free trial)
open.substack.com/pub/samf/p/t...
Reposted by James Hole
There's alot to unpack here but the very high proportions on the liberal left that would vote against Reform suggests that if Reform grow throughout the parliament some of those voters lost to Greens might vote for Labour where they are challenged by Reform.
September 17, 2024 at 9:05 AM
There's alot to unpack here but the very high proportions on the liberal left that would vote against Reform suggests that if Reform grow throughout the parliament some of those voters lost to Greens might vote for Labour where they are challenged by Reform.
Reposted by James Hole
Flow of the vote, 2019-24, revised using @britishelectionstudy.com data.
Only a few minor alterations, but the shifts are still striking - without even getting onto non voters and demographic churn, over 10m voters voted for a different party in 2024 than in 2019.
Only a few minor alterations, but the shifts are still striking - without even getting onto non voters and demographic churn, over 10m voters voted for a different party in 2024 than in 2019.
September 15, 2024 at 5:11 PM
Flow of the vote, 2019-24, revised using @britishelectionstudy.com data.
Only a few minor alterations, but the shifts are still striking - without even getting onto non voters and demographic churn, over 10m voters voted for a different party in 2024 than in 2019.
Only a few minor alterations, but the shifts are still striking - without even getting onto non voters and demographic churn, over 10m voters voted for a different party in 2024 than in 2019.
Reposted by James Hole
Via the newly released @britishelectionstudy.com data, one in eight voters really preferred another party at the general election. When adjusting for who they really favoured, we can create a 'sincere' vote share:
Lab: 31% (-4, vs actual)
Con: 23% (-1)
Ref: 16% (+1)
LD: 12% (-1)
Grn: 11% (+4)
Lab: 31% (-4, vs actual)
Con: 23% (-1)
Ref: 16% (+1)
LD: 12% (-1)
Grn: 11% (+4)
September 15, 2024 at 7:33 AM
Via the newly released @britishelectionstudy.com data, one in eight voters really preferred another party at the general election. When adjusting for who they really favoured, we can create a 'sincere' vote share:
Lab: 31% (-4, vs actual)
Con: 23% (-1)
Ref: 16% (+1)
LD: 12% (-1)
Grn: 11% (+4)
Lab: 31% (-4, vs actual)
Con: 23% (-1)
Ref: 16% (+1)
LD: 12% (-1)
Grn: 11% (+4)
Reposted by James Hole
Thoughtful ‘myth-busting’ AI article. This is a useful and succinct summary of LLM output. Suspect I may end up quoting it.
September 12, 2024 at 8:58 AM
Thoughtful ‘myth-busting’ AI article. This is a useful and succinct summary of LLM output. Suspect I may end up quoting it.
Reposted by James Hole
File under 'people are complicated'.
"...the demographic characteristics of people who are opposed to action on specific inequalities varies quite considerably."
Another good read from the @fairness.bsky.social
open.substack.com/pub/faircomm...
"...the demographic characteristics of people who are opposed to action on specific inequalities varies quite considerably."
Another good read from the @fairness.bsky.social
open.substack.com/pub/faircomm...
Who are the 15% who aren't bothered about inequality?
Our polling finds a fairly consistent 15% who are unconcerned about inequality and oppose action to tackle it. Who are these people, and are they always the same people or do they vary by issue?
open.substack.com
September 9, 2024 at 7:47 AM
File under 'people are complicated'.
"...the demographic characteristics of people who are opposed to action on specific inequalities varies quite considerably."
Another good read from the @fairness.bsky.social
open.substack.com/pub/faircomm...
"...the demographic characteristics of people who are opposed to action on specific inequalities varies quite considerably."
Another good read from the @fairness.bsky.social
open.substack.com/pub/faircomm...
Reposted by James Hole
More good stuff from the Darzi review being pre-briefed. He is absolutely right that the NHS is “undercapitalised” compared with other health systems
Since 1970, there have only been 2 years when we've invested more than the OECD average
www.thetimes.com/uk/healthcar...
Since 1970, there have only been 2 years when we've invested more than the OECD average
www.thetimes.com/uk/healthcar...
September 9, 2024 at 7:39 AM
More good stuff from the Darzi review being pre-briefed. He is absolutely right that the NHS is “undercapitalised” compared with other health systems
Since 1970, there have only been 2 years when we've invested more than the OECD average
www.thetimes.com/uk/healthcar...
Since 1970, there have only been 2 years when we've invested more than the OECD average
www.thetimes.com/uk/healthcar...