Jack Bailey
@PoliSciJack
Followers
4,152
Following
1,411
Media
825
Statuses
12,597
Political Scientist at the University of Manchester
Manchester, UK
Joined February 2009
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
SEVENTEEN
• 1097411 Tweets
Taylor
• 223357 Tweets
FELIP TAKES ON BBPHSTAGE
• 100582 Tweets
The Next Prince Q9
• 90992 Tweets
#TimnasDay
• 82056 Tweets
Carat
• 76435 Tweets
オーストラリア
• 63348 Tweets
#サッカー日本代表
• 49171 Tweets
Tattvadarshi Sant Rampal Ji
• 48430 Tweets
BBPH ANN1V WITH BINI
• 42509 Tweets
#あのクズ
• 41130 Tweets
Asnawi
• 37624 Tweets
オウンゴール
• 30728 Tweets
最終予選
• 28076 Tweets
ミンギュ
• 27182 Tweets
誹謗中傷
• 26601 Tweets
Zaplana
• 26040 Tweets
ジョシュア
• 20624 Tweets
ドギョム
• 20075 Tweets
スングァン
• 19643 Tweets
نعيم قاسم
• 19274 Tweets
Witan
• 17482 Tweets
كوريا
• 16619 Tweets
Arhan
• 14969 Tweets
中村敬斗
• 12469 Tweets
Tuchel
• 11549 Tweets
報道ステーション
• 10833 Tweets
Pinned Tweet
📊 We have a new paper
@Res_Pol
.
Key point: LLMs perform almost as well as trained humans at coding open text (here the most important political issue). But they cost a lot less and are faster. They also out-perform prevailing supervised methods too.
1
44
108
Rishi Sunak built up all that political capital then blew it all over this crisis. Wouldn't be surprised is Starmer soon has the highest approval of any politician on this tracker.
40
172
829
People are now ≈0.3 points happier on a 10-point scale than since we started recording it in the aftermath of the greatest financial crisis in a generation.
Truly, we live in a golden age.
13
73
386
Ok, so now the constituencies that pollsters need to oversample for their MRP polls are:
- Brighton Pavilion
- Bristol Central
- Islington North
- Clacton
Any others?
62
15
353
Not spoken up until now, but I was in charge of organising and managing the Germaine Greer event at Cardiff University in 2015.
I can assure everyone that it most definitely went ahead because the whole thing still haunts my dreams.
@epkaufm
Your question reads “the university should have stepped in and ensured Germaine Greer could speak”. This is in fact what the university did. The way the case is presented strongly implied it did not. Is this not deception? Is it not misrepresenting Cardiff university?
4
17
173
11
103
323
Compare public sentiment towards Sunak when he became PM (left) and now (right). Shows just how costly it can be to focus on second order culture war issues rather than delivery.
(Stolen from
@drjennings
who found the older word cloud)
7
94
269
😎🤘
Big thanks my supervisors Ed Fieldhouse,
@MartaCuanti
,
@jon_mellon
for making it possible!
Congratulations to Dr
@PoliSciJack
who passed his PhD viva with flying colours, and no corrections, this afternoon. An outstanding piece of work from an outstanding young scholar - original contributions on important questions. And a pleasure to read too! Well done Dr Bailey. 🍾
7
7
173
72
8
220
There have been quite a few conflicting polls over the past few days. Here's where our model thinks things are as of this morning ✌️
Con: 38% (35% to 40%)
Lab: 37% (34% to 39%)
Lib: 10% (9% to 12%)
Con-Lab lead: 1pt (-2pt to 5pt)
16
67
219
🚨SURVEY NERDS! 🚨
I have a new piece out in Public Opinion Quarterly that I think you might like (
@AAPOR
). It's all about how political surveys can bias the economic perceptions that people report.
Pre-print link:
5
45
182
The polls are now so extreme that there would be a genuine constitutional crisis if they occurred at an election. Can't have 1 Lib Dem MP as the opposition. It would be like Manchester City Council, but country-wide.
NEW: Westminster Voting Intention poll (20 Oct):
🔴 LAB: 53% (= from 12 Oct)
🔵 CON: 14% (-5)
🟠 LDM: 11% (+3)
🟢 GRN: 6% (=)
🟡 SNP: 5% (-1)
Full tables:
156
1K
3K
19
22
152
Really boring seeing people compare vote counts. Yes, Starmer's is lower than Corbyn's. But it's *broader* too. It would have been easy to win votes in safe seats, but that doesn't win seats.
You work with the system you've got, that's what Starmer did, and that's why he won.
30
38
153
Did they change the text layout under MEDIUM and HIGH to obscure the fact that they say *exactly* the same thing?
Here’s how the rules on pubs and bars differ between Local COVID Alert Levels.
More information:
446
128
545
1
34
134
Labour's lead over the Conservatives is now around 11 percentage points.
To put that figure in context, it's the highest lead for Labour since *2002*, two decades ago!
4
37
125
In personal news, I start a new role on the British Election Study as a Research Associate this week. Really couldn’t be happier about it. Can’t wait to work with a great team and — most importantly — to help make possible all of your excellent research!
10
0
123
Just going to piggyback on Lotte’s tweet and mention that I will *also* be starting as a Lecturer in Quantitative Politics tomorrow at
@UoMPolitics
. Couldn’t be happier about it and am thrilled not to have to start the job alone 😅
Since I’m in (v rainy) Manchester for the day, it feels like an ideal time to announce that, from tomorrow, I am officially coming back to academia!
Really excited to be joining a fab group of people at
@UoMPolitics
@DemElect
as a Lecturer in Quantitative Political Science 🎉
21
5
158
20
4
124
And, for the record, the University never, ever -- not once! -- considered cancelling the event. Greer threatened to pull out of her own accord, but didn't.
On top of that, the crowd of protesters at the event itself included maybe 10 people.
1
17
115
Mark Reckless (former Conservative MP turned UKIP MP turned UKIP MS turned Conservative MS turned Brexit MS turned Abolish the Assembly MS) is probably going to lose his seat next month and if there's a better case of nominative determinism I don't want to hear it.
5
19
112
UK pollsters: pool all of your micro-data into a single repository and sell subscriptions to universities, I beg of you. There's so much great micro data out there that just gets lost forever and it's a travesty.
5
6
108
I have the Conservatives on ≈41% and Labour on ≈38%.
A reminder: even if Labour aren't 20 points ahead, they still stand to gain around 40 seats and deny the Tories a majority at their current polling level.
11
37
107
Stats nerd/Bayesian Twitter: What are some of the coolest non-linear models that you've seen implemented in
@paulbuerkner
's {brms}?
Three of my favourites 👇
2
14
109
I modelled vote share over time. The Opinium poll everyone was holding up as evidence of X/Y/Z yesterday is within the range we'd expect given past polling.
You're opining about noise.
3
49
107
🚨📊 A new paper of mine, "Political Surveys Bias Self-Reported Economic Perceptions", was just accepted at Public Opinion Quarterly (
@AAPOR
).
The preprint is available now at
@APSA_Preprints
:
Let me break it down for you...👇
9
25
107
If you showed this to someone in December 2019 they'd question if we ever left the EU
NEW: First Westminster voting intention since Farage announced his candidacy.
LAB 43 (-4)
CON 23 (-1)
LD 9 (-2)
GRN 5 (+2)
RFM 15 (+7)
SNP 3 (-)
OTH 3 (-1)
F/w 5th - 6th June. Changes vs. 27th May 2024. 1,056 adults in Great Britain.
57
175
549
4
19
102
Updated British poll of polls, 8 November 2021:
Con: 37% (34% to 39%)
Lab: 37% (34% to 39%)
Lib: 11% (9% to 12%)
Con-Lab lead: 0pt (-4pt to 4pt)
4
35
99
Folks, it looks like it’s really happening. Look at that drop in daily cases! We’ll be on 10% having had the first dose before you know it too.
7
10
98
This is my takeaway too. Random probability samples are "gold standard" when it comes to survey research. That it puts Reform so much lower than internet-based polling is informative in my view.
2
26
97
📈 Thread: 2019 to 2024 Polling Post-Mortem 📉
There have been a few attempts to rate pollster accuracy over the last parliament. The problem is that random error & fieldwork dates can affect their results. So I've done some modelling to get around this.
Results below 👇
3
26
99
@ProfSobolewska
The way that the "NATO is expanding eastward" logic infantilises Eastern Europeans is vomit-inducing
2
4
93
I've heard government ministers refer to Rwanda as a democracy. This is not true.
According to the best data we have -- the
@vdeminstitute
electoral democracy index -- Rwanda is not a democracy and never has been.
6
35
94
Rereading Downs’ An Economic Theory of Democracy. Immediately understand why some people hate this stuff.
5
4
91
Today's PMQs really highlights the beautiful absurdity of our system.
Labour/SNP MPs: RESIGN!
Conservative MPs: But what about the washing machines?
0
14
86
Cool plot that immediately jumped out at me: the runner up map of Britain according to the recent YouGov MRP release.
15
40
83
Mini-thread:
Here's a plot you're all probably used to seeing. It's taken from Wikipedia's British opinion polling page. It clearly shows the Conservatives heading ⬆️ and Labour heading ⬇️...
5
25
79
Compared to the absolute bodying that Sunak got, this is pretty positive. Obvious line to take too: "It's clear that the public want to know what I stand for and what Labour stand for and I'm going to set that out in my conference speech".
Keir Starmer’s turn for the BBC word cloud treatment - a mixed bag but “don’t know”, “nothing” and “not sure” indicate there’s more work for Labour to do.
175
90
532
3
14
80
Remember
@stephenkb
saying that he thought the Conservatives had been more effective during the pandemic due to their staff not working from home, unlike Labour. Think I'd agree, but funny how it's come back to bite them.
3
5
77
@johnwrabin
@JadeSymonds
@marcusbrig
They didn’t run out of money. It’s literally impossible for a sovereign government to run out of money. Those notes are a running tradition from one office-holder to the next, but which Cameron used to his own ends. It was naive, but didn’t actually mean there was no money.
2
0
66
Today’s one of those rare days where Britain looks like all the period dramas pretend it looks.
2
5
76
The Conservatives should be really, really worried. The current context (partly by their own design) is now so dire that even to avoid a wipeout will be a major achievement.
Things that will be happening over the next 5 months.
* More tax rises
* Spending cuts
* 2.6m households with higher mortgage payments
* Public sector strikes
* Possible blackouts
* Energy bill help reduces
That's before you get to the NHS crisis.
69
384
2K
2
15
73
This poll puts independence support at:
Yes 46%
No 43%
DK/etc 12%
Its sample is 1019. Assuming random sampling error only (a conservative assumption), these figures have a MoE of ~3%. I.e.:
Yes 43-49%
No 40%-46%
Uncertainty matters to polls. Even if it ruins your headline.
🏴 A majority of people in Scotland now back independence, according to the latest
@LordAshcroft
poll:
Yes: 52%
No: 48%
🗳 Almost half back an independence referendum within two years.
33
176
573
4
29
73
This article is an *excellent* example of just how important political science is for our democracy. Powerful and clear. The conclusion says it all.
Read
@chrishanretty
latest article for PQ "The Pork Barrel Politics of the Towns Fund" (no paywall)
#porkbarrelpolitics
#TownsFund
#MinistryofHousing
#Nolanprinciples
1
36
49
3
30
69
Just over one week left in Austin.
Did I read all the books I brought with me? No.
Did I finish all the writing I wanted to do? Also no.
But did I eat over 100 tacos? Yes, yes I did.
5
0
70
In all honesty, the Cardiff event was just badly timed and the (very light!) student hostility coincided with recent events in the US. To be honest, I think that it was just a good story and the media ran with it.
1
4
70
Self-isolating during this spate of nice weather has been miserable. But I just completed a first draft of my PhD thesis, so things could be worse 🙌
2
0
67
For my sins, I rewatched
#BBCDebate
and made a network of every time someone interrupted someone else. TLDR: Little parties shout a lot.
6
60
66
🧵 A little thread on why you need to take a long-view of the polls right now. First up, here's today's poll of polls.
Con: 38% (36%-41%)
Lab: 38% (36%-40%)
LD: 7% (7%-8%)
1
30
66
Student hostility started when a student rep tweeted out that she opposed Greer speaking due to her views and past comments on trans people. Actual student hostility on campus was limited and mostly occurred online/in the media.
1
5
63
@Mishi_Mo
@robfordmancs
As I make clear, there was no serious attempt from within the university. A very small number of students objected and then the media pouncesd a it was a good story. And I'm being flippant -- I'm not actually traumatised! It was stressful and stupid, but not traumatic.
1
2
60
I have my viva on Thursday and am currently preparing. If anyone who's been through the experience has any tips, I'd be grateful to receive them!
32
5
66
@PhilipPullman
Because the political identities and allegiances that people form are often not due to reason but socialisation and other sociological phenomena.
3
3
64
Mad story: I was walking my dog yesterday and dreamt up a new type of regression. When I got home, I scripted it up and got it working in ≈ 20 mins.
Would love some feedback from my more techy followers.
3
3
64
New post on my blog:
Since there's no official election repository, historic election results are becoming more and more uncertain. We need to do something about it ASAP before time runs out ⏳
9
18
65
I think we all need to stop for a moment, be present, and reflect on how mad these numbers are.
We're months from an election and the governing party is skirting around *20%*. Once you add back in DKs, you're looking at maybe only 1/10 people saying they support them.
Lowest Conservative % since Sunak became PM.
One point above their lowest under Truss.
Westminster VI (21 April):
Labour 43% (-1)
Conservative 20% (-2)
Reform 14% (-1)
Liberal Democrat 12% (+3)
Green 6% (–)
SNP 3% (–)
Other 1% (–)
Changes +/- 14 April
45
248
602
2
11
62
A take on academia’s apparent left-wing bias:
Many comments about
@epkaufm
’s survey point to it being non-representative, too small, etc. This might be true, but misses the point: even a perfect survey of *current* academics does not give us the inferences we need.
3
15
57
Amazing that Labour and the Conservatives have managed to engineer a situation where voters are *more* likely to think that taxes will go up under a Conservative government than a Labour one.
Here’s another way of looking at it, the difference in public expectations on waiting lists depending on who wins is massive, whereas expectations around interest rates and inflation are very similar between the two parties.
2
5
13
2
14
62
Idea for a survey experiment:
- Assign Brits to one of 2 conditions at random
- The first get "Around 1 in 1,000 people has died"
- The second "Around 66,000 people have died"
- See how they evaluate the Government's performance
2
8
60
@northofcam
@robfordmancs
It was 5 years ago and the context was almost exactly the same, so I imagine almost exactly the same would have happened too.
3
0
56
@paulwaugh
Hi Paul, as someone who knows a little about what this means, I can confirm that the text here includes no information whatsoever on how this person managed to overcome the sampling bias inherent in their survey.
4
6
58
@Samfr
Reminds me of Brecht:
...the people
Had squandered the confidence of the government ... Would it not in that case
Be simpler for the government
To dissolve the people
And elect another?
1
4
55
Labour strategists: Can't be caught out having illegal fun if you're no fun at all
1
3
58
Rishi’s right to be concerned: they didn’t remove stop words or stem properly 😱
This word cloud for what people think of Rishi Sunak is brutal…
471
2K
12K
3
3
57
I remember travelling from Cardiff to my home town and getting an urgent email from my boss saying that Greer was on Newsnight. I thought that the world had gone crazy.
1
2
53
If you’re going to criticise it on GB News, you should criticise it on ITV too. Bluntly, politicians and former politicians probably shouldn’t front TV news full stop.
'Beyond Parody': ITV Criticised After Home Secretary Yvette Cooper Interviewed By Her Husband Ed Balls
9
9
16
2
7
54
Keep in mind that the sample for this question is 323. So a conservative estimate of the margin of error on the Con figure is about 5.5 percentage points.
Batley & Spen, constituency voting intention:
CON: 47% (+11)
LAB: 41% (-2)
WPGB: 6% (+6)
LDEM: 3% (-2)
No Local Ind Grp (-12) and Brex (-3) as prev.
via
@Survation
, 09 - 17 Jun
Chgs. w/ GE2019
342
488
2K
7
12
53
These simulations are all well and good, but have 1 problem: the election was fought under FPTP. If it was fought under PR, the strategies would have been different and so would the vote shares.
The fact is that we can’t really know what outcome PR would have produced:
3
10
54
Unless there are a secret Labour parties we don't know about, but I doubt it very much given that no one in Labour knows how to have a good time
4
3
54
@Mishi_Mo
@robfordmancs
And, bluntly, almost all of the heat came from people desperate to find some evidence of on-campus threats to free speech which didn't exist or from media organisations reporting on the event as though such a conflict existed at all.
1
1
50
Someone needs to offer an intro to psephological course for MPs because this is insane
This section is very uh.
Put it this way. These are not average seats. If you're only *just* coming out 'on top' in two seats you already hold, one of which was basically in the safest category of seat what do you think that implies about other seats.
25
44
172
4
9
53
If Johnson was still PM, the outcome would likely be the same. But something more interesting is that claims like this reveal how many people (and many politicians) make counterfactuals: by comparing the present vs the past and not the present vs an alternate present.
Thank God for those clever-clog ‘Tory grandees’ who got rid of Boris. Dodged a bullet there didn’t they! Genius 🤣
453
258
1K
7
6
53
I have some new work out
@ElectoralStdies
.
TLDR: Party identification moderates economic learning and partisans react to the economy in ways that flatters their party.
Free access until 1 Nov with this link, so click away.
Let me break it down 👇
4
18
52
Angela Merkel has been chancellor since 2005 and is still around 20pts clear of her nearest rival.
That’s the same year George W. Bush was inaugurated for his 2nd term...
Germany, YouGov poll:
CDU/CSU-EPP: 35%
GRÜNE-G/EFA: 18%
SPD-S&D: 15%
AfD-ID: 12%
LINKE-LEFT: 9%
FDP: 6% (+1)
+/- vs. 19-21 Sep
Fieldwork: 22-26 October 2020
Sample size: 1,656
➤
4
9
58
4
11
49
@GoodwinMJ
Why would you ever expect Twitter to be representative of the country? And even if it was representative of the country no one would have a representative Twitter feed. That's the point of Twitter!
2
1
47
Also a totally insane strategy given that it seems to be based on a single (very slim!) by election win in London.
A reminder: Just because people seem to think that ULEZ won it for the Conservatives doesn't mean that it actually won it for the Conservatives!
Some in the Conservative party seem to have concluded that ULEZ/Uxbridge gives an electoral green light to campaigning on every obscure gripe the Very Online right has about urban planning. I'm dubious there are votes in this - its simply not what voters are focussed on
35
81
514
4
7
51
Boris Johnson's constituency is now *much* more competitive than YouGov's last MRP predictions. It'll likely still go his way but even so...
The banter heuristic says that we're in for a wild ride. A Conservative majority of 80+ seats, but Johnson isn't one of them.
7
13
49
The plot below shows the resulting values. According to the model estimates, Verian came in first *and* second, since Kantar has since rebranded as Verian. The distinction here is useful, however, since polls under the Varian brand used a random probability sample.
4
7
49
@PhilipPullman
Likewise, people often do not have their own reasoned positions on things, so use their party identification as a heuristic or engage in some behaviour (voting for Brexit) that has implications for their own self-conception.
3
0
46
@SeanDEhrlich
Use RStudio. It's great and easy to get a hang of. This book is great too. Written by someone at RStudio, free, and comprehensive. Good luck!
3
3
46
Spare a thought for those of us who work out of one-bedroom flats with no dedicated office space. Spoiler: it's not a productive way to work.
7
1
44
I’m going to start a DM group of people using
@mcmc_stan
to do political science research.
The intention: to offer each other advice/share interesting and useful stuff we find.
If you use Stan and want me to include you, like this and I’ll invite you when I get round to it.
6
5
44
@profmusgrave
Brit here. Military would almost never board a civilian aircraft wearing uniform or wear uniform if they weren’t in the middle of some operation. I’d probably be worried if military personnel in uniform got on my flight.
2
0
44
Over two years since the first lockdown, I'm still amazed how much of an impact it continues to have on me: my brain feels like Swiss cheese
1
1
43
I’m reluctant to say it because it’s a complete nightmare for me, and my colleagues will hate it too, but I fear we need a general election
9
5
45
According to our estimates, Labour hold a lead over the Conservatives of around 3 percentage points (0pp to 7pp).
3
16
43
Of everything I've seen, this is the thing that best explains why this MRP looks comparatively rosy for the Conservative.
If this assumption isn't right, things could be worse!
This MRP model also probabilistically works out where the “missing” VIs from the sample data (ie, the “Don’t Knows”) will go, using information about similar *types* of voters who *do* have a VI.
That means a boost for the Tories, as most DKs now “look like” Tory voters.
3
6
27
3
12
43
Ricotta, potato, and rosemary pan pizza with caramelised onion, truffle pesto, and garlic honey 😋
#poliscicooks
2
0
44
@GoodwinMJ
The BBC is probably no less partisan than it's ever been, it's just that people see it as more partisan because they themselves are more partisan. Plus, there are now more issues to be partisan about too (e.g. Scottish independence, Brexit).
4
2
43
The Conservatives and Labour both appear to be gaining ground. Lib Dems now down to 14% and Brexit just 6%.
UK Poll of Polls, 20 Nov 2019
Con: 42% (38%-45%)
Lab: 30% (27%-33%)
LD: 14% (12%-16%)
Brx: 6% (5%-8%)
Nat: 4% (3%-5%)
Grn: 3% (3%-4%)
13
46
40
@blain321
@fish_in_a_hat
@DChildEsq
@cinnamonbear14b
@davidandvinny
@johnredwood
The money we give is recompense for the programmes we pledged to fund (inc. Brits’ pensions). And it’s “ça suffit”.
2
1
40
It would be great if the
@ESRC
made it compulsory for any data collection projects that received funding to make their data available even if they weren't used in published research. There are huge and useful datasets out there just languishing on hard-drives.
3
4
43