🚨 With 8 days left until polling day, this is the lowest Conservative share we’ve recorded
📈At the same time, Starmer has doubled his lead over Sunak on the best PM question since the start of the campaign
NEW: First MRP Since Farage’s return.
Labour on course for 262 seat majority.
LAB 456
CON 72
LD 56
SNP 37
RFM 7
PC 2
GRN 1
42,269 interviews conducted online and on the phone
Fwk 31 May - 13 June
Conducted on behalf of
@BestForBritain
.
@VasilSurvation
and
@JackSurvation
NEW MRP: Labour 99% Certain To Win More Seats Than in 1997
Labour on Course to Win 484 seats.
The Conservatives and Liberal Democrats are in a close race to form the official opposition.
Probabilistic seat count:
LAB 484
CON 64
LD 61
SNP 10
RFM 7
PC 3
GRN 3
34,558 interviews
NEW: Our first MRP of the 2024 General Election
Labour Set for Record Breaking Majority
LAB 487
CON 71
LD 43
SNP 26
RFM 3
PC 2
30,044 interviews conducted online and on the phone
Fwk 22 May - 2 June
Conducted on behalf of
@BestForBritain
@VasilSurvation
looks at the findings
Elphicke's associations with the right on social issues are well-documented, but we found that she is also to the right of the median Tory MP on *economic issues*.
Just 10 MPs are more right wing on economic issues, including Truss and Braverman.
Survation MRP Update
👉 Reform are currently leading in 7 seats
👉 Reform are also currently performing better than the Conservatives in 59 seats
👉 BUT their largest lead at the moment is 8% (North West Norfolk)
NEW: First MRP Since Farage’s return.
Labour on course for 262 seat majority.
LAB 456
CON 72
LD 56
SNP 37
RFM 7
PC 2
GRN 1
42,269 interviews conducted online and on the phone
Fwk 31 May - 13 June
Conducted on behalf of
@BestForBritain
.
@VasilSurvation
and
@JackSurvation
National service policy is evidently an attempt to reclaim Reform voters, but it’s not a panacea for the Tories.
Our MRP for
@BestForBritain
in March revealed that even in the ideal scenario where Reform UK doesn’t stand any candidates, the Conservatives would win only 150 seats
NEW: Our first MRP of the 2024 General Election
Labour Set for Record Breaking Majority
LAB 487
CON 71
LD 43
SNP 26
RFM 3
PC 2
30,044 interviews conducted online and on the phone
Fwk 22 May - 2 June
Conducted on behalf of
@BestForBritain
@VasilSurvation
looks at the findings
NEW: Penultimate MRP Update.
Labour on Course to Win 470 Seats and Become the Largest Party in Scotland.
Probabilistic seat count:
LAB 470
CON 85
LD 56
SNP 12
RFM 4
PC 3
GRN 2
23,364 interviews conducted online and on the telephone
Fwk 15th - 27th June
Key point from
@chrishanretty
👇
If you *only* use historic data on elections under FPTP and you expect a Labour lead of around 19 points, you should expect them to win 68% of the seats, or 442 seats.
Our first poll of Scotland in the election campaign places Labour 4 points ahead of the SNP.
Labour are attracting voters from all parties. Current support is made up of:
21% 2019 SNP voters
17% Con
33% LibDem
47% Other
(usual caveats about smaller sample cross breaks apply)
Labour lead the pack among voters in Scotland with little over five weeks until the country goes to the polls.
36% of Scots are set to back Labour in a Westminster vote.
The SNP could expect to win 32% of the vote — six points down on where they stood just two months ago.
Reform’s performance will be the key difference between a bad night and a total wipeout for the Tories.
Reform are currently more popular than the Tories among 18-24 year olds and as popular among the 45-54 group.
Labour are in the lead among all but 65+
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.
Bella couldn’t wait for the 10pm exit poll so decided to conduct her own by sniffing everyone who went into the polling station 🐶
@VotingDogs
#dogsatpollingstations
NEW MRP: Labour 99% Certain To Win More Seats Than in 1997
Labour on Course to Win 484 seats.
The Conservatives and Liberal Democrats are in a close race to form the official opposition.
Probabilistic seat count:
LAB 484
CON 64
LD 61
SNP 10
RFM 7
PC 3
GRN 3
34,558 interviews
🧵What’s going on with the SNP’s lead in Scotland?
SNP currently at +3 over LAB - lowest we’ve recorded since 2020
IndyRef2 attitudes stable - Yes 47, No 53
Scots want more devolution - 41%+
BUT Unconvinced about economic benefits of independence - 43%+
1/11
Key point here is that there are bound to be some outliers but the main benefit of MRP is to get constituency level estimates (for all constituencies) with significantly lower sample sizes than individual seat polls.
Example - Godalming and Ash: Phone Poll vs MRP
Is the Conservative Party facing its worst defeat in history?
@VasilSurvation
breaks down what our latest MRP analysis for
@BestForBritain
does and doesn’t tell us in this article:
Heads up for anyone using new constituency codes - ONS have updated their codes for 5 constituencies in Scotland.
Ours are all up to date but it might cause some discrepancies when comparing MRPs.
Final (hopefully) names and codes here:
Since we are seeing many MRPs on the horizon, it would be very useful (and good practice) to see them accompanied by a measure of uncertainty in the seat estimates.
We will be publishing high and low 95% credible intervals along with win probabilities for each seat estimate.
New ward level voter ID MRP in London shows considerable variation within Boroughs e.g.,
in Tower Hamlets:
Bow West - 8.1%
Bethnal Green - 3.3%
Mile End - 7.2%
New: 5% of registered voters in London lack a valid form of photo ID
Ward level estimates show that this the highest in Becontree - 11% (Barking and Dagenham) and lowest in Bloomsbury - 0.5% (Camden).
Final Survation Poll of the 2024 General Election:
18 point Labour lead
Labour: 37.6%
Conservative: 19.9%
Reform UK: 17.0%
Liberal Democrats: 12.1%
Green Party: 7.2%
Scottish National Party: 3.0%
Plaid Cymru: 0.6%
Other: 2.4%
Based on telephone interviews of 1,679 respondents
@MattMcMaster24
@pimlicat
@DamianSurvation
MRP produces different estimates outside of election periods because it uses data from people who tend to act differently during elections.
Tactical vote recommendations are based on MRP with data collected during the campaign, not several months before polling day.
I definitely wouldn't trust the London Mayor VI results here. Looking at the tables, something's definitely gone wrong with the weighting:
-> 2021 Shaun Bailey voters are inflated by 6% in Mayoral tables vs others.
-> 5% higher share for 2019 Tory voters...🙅♂️
🚨Dashboard Klaxon🧵
We worked with our partners at Lodestone Communications to build a GE 2024 dashboard, which has some cool features.
The initial focus is on Labour vs Conservative battles, but there's MRP data and projections for all seats in GB:
@charliebuffet99
Hi Charlie, MRP doesn’t deal very well with very specific local factors. We like being transparent about our method and data so we don’t make manual adjustments to our estimates.
Some insights from Mid Beds...
👉 Tactical voting will be key
👉 19% of all voters are voting against the Tories
👉 Con would need to mobilise rather than convince
👉 42% of undecideds are certain to vote
👉 A lot will change between now and polling day
More here:
Having looked a further at
@Survation
's tables, can I just ask, please, *every* pollster, follow Survation on this.
Constituency codes, hi, lo and mid vote shares, win probability's current winner&2nd place, 2019 winner and 2nd place
It's fantastic, much more to learn from it.
@WTarbiat
Margin of error is a lot higher with such small samples. Lib Dems have a relatively flat level of support across age groups. I would expect them to be closer to 8-10% with a larger sample.
As the releases are a bit confusing, here's the new Martin Baxter MRP. 2 Versions:
Seat Projection Standard Method:
LAB 493
CON 72
LD 39
SNP 22
PC 4
GRE 2
REF 0
Seat Projection Tactical Voting Adjustment
LAB 476
CON 66
LD 59
SNP 26
PC 3
GRE 2
REF 0
Vote Share Figures Implied
In the last 3 months SNP is down -3 points (Westminster vote)
40 in April
38 in May
37 in June
Some would’ve expected a more drastic drop, given recent events
2/11
Given the above, it is unlikely that voters will make their minds up with independence at the forefront of decision making.
More (44% vs 35%) believe an independent Scotland’s economy would be worse off, and 43% vs 31% believe the same for CoL.
8/11
@thesourwaitress
@Survation
The estimates here actually seem to be from MRP analysis (old boundaries) on data we collected b/w 10-17 Feb for
@38degrees
, here's the link:
NEW: First Westminster voting intention from the campaign period.
23 point Labour lead.
LAB 47 (-1)
CON 24 (-3)
LD 11 (+3)
GRN 3 (+1)
RFM 8 (-)
SNP 3 (-)
OTH 4 (-)
F/w 24th - 27th May. Changes vs. 22nd May 2024.
@herculeanstable
@BestForBritain
I think all pollsters who use MRP do well with reporting their methodology and it's limitations but we can't control how third parties report it.
We love doing constituency polls, would be more than happy to assist if you'd like to commission one!
BUT LAB is currently attracting the lowest share of 2019 voters from the SNP. Instead:
25% of 2019 LD voters
23% of OTH parties
20% of CON
16% of SNP
5/11
Labour in the same period is up +2 points
32 in April
31 in May
34 in June
At first glance, it seems that Labour might be gaining voters directly from the SNP…
3/11
@ABLPoli
I haven't seen the tables but this usually happens if you set a hard turnout threshold (say 7+) and hard-code others' weights to 0, then rebase. Personally not a fan
All in all, the SNP continues to be in the lead, despite the decreasing margin.
It suggests good levels of stickiness with their core base, though in our *UK* party identity polling, they are the party with the weakest supporters in terms of strength.
10/11
Thus, the next GE is more likely to be fought on arguments for devolution than independence. With Labour on track to become the largest party in Westminster, their stance on devolution might be the key to swaying Scottish voters.
Full analysis here:
11/11
At the same time, Scots are on average in favour of further devolution on matters of:
Cost of living policies 46%
Energy security 42%
Employment law 41%
9/11
Thus, it should be worrying for the SNP that LAB is making gains across all parties.
Combine this with stable attitudes on independence (No - 53, Yes - 47), and you get an idea about the issues over which the next election could be decided in Scotland.
7/11
Very interesting paper - affective polarization on immigration increased in homogenous groups (of Lab and Con voters), but decreased significantly in mixed groups
@herculeanstable
@BestForBritain
The point is to understand how national vote shares could translate into seats under a FPTP system. Here is how the method works:
We carried out a similar exercise last year (on new boundaries) and found very similar results.
(this is for the new seat - Wellingborough and Rushden)
@robjeffecology
@Survation
Hi Rob, we use a full ballot prompt method and don’t make adjustments (reallocation, modelling, etc) for undecided.
Voting intention is the first non-demographic question in every survey.
From the Redfield London poll.
A person in London is apparently *less* likely to be a Westminster Conservative voter if they are in their 50's and 60's than if they are in their 20's and 30's.
For Labour voters, there's no age effect at all on their vote from 25-65.
🤷♂️
@ramshackle78
It's a relatively even increase around 5-8% across all age groups (in line with national increase), though bear in mind the higher margins on cross breaks.
Mostly indicates a "convention bounce" as
@benwansell
points out:
@Samfr
Is nobody in the country aware of ‘convention bounces’? A party gets a lot of news when something big happens. Like Farage standing as MP. And that brings a lot of attention and a surge in intended vote. Typically it subsides. Not always. But typically.
@livvyjohn
@scotfax
@Survation
@BestForBritain
13 of those are with margin < 5%, and in the probabilities tab, only 20 SNP wins are with chances higher than 3 in 4, so the difference between 40 and 20-25 seats is very small. 2014 "No" voters being largely split between Lab/Con contributes to that
@sundersays
@TomHCalver
@Survation
Not very promising for them if the share of current Undecided voters who would *consider* voting Con is roughly the same as the share of Green voters…
From the same poll:
@herculeanstable
@BestForBritain
All companies that produce MRP have a set number of minimum interviews per constituency and I’m assuming many would be oversampling specific seats like Clacton.
Is this likely to pick up all local factors that a constituency poll would? No, but it’s the next best thing.
@AdrianBeavis1
@MattRanson
It’s very much in line with the one published yesterday morning. All changes are within MOE, main differences are that this is GB not UK and it has shorter/more recent fwk period.
No plans to make implied shares from MRP the main VI 😀