General article(s) on Polls

[greenchristian]

May 27, 2024 12:33:15 GMT greenrichard said:

May 27, 2024 10:47:31 GMT johng said:

You have to wonder.

There's just no evidence from by elections or local elections that Reform are on up to 14 points.

Local elections aren’t a good predictor of national votes for minor parties. I think this is especially true of Reform UK as they don’t seem to value local elections in the same way as Greens and Lib Dem’s do. They don’t stand that many candidates and I don’t think they have a local organisation in place.

The reason they aren't a good predictor for minor parties has always been because minor parties perform significantly better at local elections (due to a combination of lower turnout, major party supporters being willing to lend their votes to minor parties in locals, and minor party supporters being willing to lend their votes to major parties in general elections). If Reform does perform very differently at the GE to how it does in LEs then it will be for very different reasons to those that hold for other minor parties.

[nodealbrexiteer]

Not an article as such but Peter Hitchens has put out a few tweets like this about the 1970 election and how the polls were wrong, arguing this election isn't over:

x.com/ClarkeMicah/status/1796220414638575759

[John Chanin]

May 31, 2024 16:45:36 GMT nodealbrexiteer said:

Not an article as such but Peter Hitchens has put out a few tweets like this about the 1970 election and how the polls were wrong, arguing this election isn't over:

[nodealbrexiteer]

May 31, 2024 17:23:53 GMT John Chanin said:

May 31, 2024 16:45:36 GMT nodealbrexiteer said:

Not an article as such but Peter Hitchens has put out a few tweets like this about the 1970 election and how the polls were wrong, arguing this election isn't over:

Sorry couldn't see your message at the bottom of your reply?

[Davıd Boothroyd]

Although their overall results are very similar, there are no fewer than 97 seats where the More in Common and YouGov MRP results have different winners.

Labour in YouGov, Conservative in MIC (31)

Aldershot
Ashford
Aylesbury
Banbury
Bridgwater
Cannock Chase
Central Devon
Chatham and Aylesford
Derbyshire Dales
Forest of Dean
Great Yarmouth
Harwich and North Essex
Hexham
Isle of Wight East
Mid Bedfordshire
Morecambe and Lunesdale
North East Hertfordshire
North East Somerset and Hanham
North Somerset
North Warwickshire and Bedworth
North West Cambridgeshire
Penrith and Solway
Reading West and Mid Berkshire
Redditch
Ribble Valley
South Basildon and East Thurrock
South Dorset
South Norfolk
Stoke-on-Trent South
Tatton
West Suffolk

Liberal Democrat in YouGov, Conservative in MIC (24)

Bicester and Woodstock
Chesham and Amersham
Chichester
Chippenham
Didcot and Wantage
Godalming and Ash
Harpenden and Berkhamsted
Honiton and Sidmouth
Newbury
North Cornwall
North Devon
North Norfolk
North Shropshire
South Cotswolds
South Devon
St Ives
Surrey Heath
Sutton and Cheam
Thornbury and Yate
Torbay
West Dorset
Woking
Wokingham
Yeovil

Labour in YouGov, Scottish National Party in MIC (21)

Ayr, Carrick and Cumnock
Central Ayrshire
East Renfrewshire
Na h-Eileanan an Iar
Kilmarnock and Loudoun
North Ayrshire and Arran
Alloa and Grangemouth
Bathgate and Linlithgow
Cumbernauld and Kirkintilloch
Dunfermline and Dollar
East Kilbride and Strathaven
Edinburgh North and Leith
Falkirk
Glasgow South
Glasgow West
Glenrothes and Mid Fife
Inverclyde and Renfrewshire West
Livingston
Paisley and Renfrewshire North
Paisley and Renfrewshire South
West Dunbartonshire

Conservative in YouGov, Labour in MIC (12)

Broxbourne
Dumfries and Galloway
Epsom and Ewell
Hertford and Stortford
Hornchurch and Upminster
Montgomeryshire and Glyndwr
Old Bexley and Sidcup
Orpington
Romford
South West Hertfordshire
Spelthorne
Sutton Coldfield

Scottish National Party in YouGov, Liberal Democrat in MIC (4)

Caithness, Sutherland and Easter Ross
Edinburgh West
Mid Dunbartonshire
North East Fife

Conservative in YouGov, Liberal Democrat in MIC (2)

Eastleigh
Wells and Mendip Hills

Conservative in YouGov, SNP in MIC (1)

Aberdeenshire North and Moray East

Green Party in YouGov, Labour in MIC (1)

Bristol Central

Labour in YouGov, Plaid Cymru in MIC (1)

Caerfyrddin

[batman]

The YouGov results make more sense by & large. It strikes me as silly to predict that Labour will gain Broxbourne yet fail to win Morecambe & Lunesdale, to give just one example. Though YouGov's predictions of SNP gains from the Lib Dems also seem pretty poor.

[evergreenadam]

Jun 3, 2024 22:17:36 GMT batman said:

The YouGov results make more sense by & large. It strikes me as silly to predict that Labour will gain Broxbourne yet fail to win Morecambe & Lunesdale, to give just one example. Though YouGov's predictions of SNP gains from the Lib Dems also seem pretty poor.

The comment made elsewhere that 2/3 of the fieldwork for MIC was carried out in April, may help to explain the differences. The launch of the Tory campaign has made things worse for them, not better!

[johnloony]

I’m looking at the opinion polls generally with a pinch of caution and an elbow of salt. The last several polls in the last few days have ll got Labour on 41%, 42%, 43%. Even 41% would normally be a fairly modest percentage for a winning party in a general election. In 1997, the polls were putting Labour on c.50%, with the exit poll on 47% and the actual result on 43.2%.

When I made my prediction in the competition about 2 weeks ago, I was guesstimating on the basis ofmassuming a Labour vote of about 39%. Even if it gets down to 37% it might still be enough for a landslide, depending on how the rest is split.

[The Bishop]

Labour support was not significantly overestimated by polls (indeed in some cases the opposite) in the 2019/2017/2010 GEs, and only modestly so in 2005. These elections are arguably of more relevance to today than 1997 - when most polling was conducted in a majorly different fashion to now anyway.

(yes I know about 2015, but the reasons for that miss were mostly about poor sampling and bad turnout estimates - not the far too lazily trotted out "shy Tories")

[hullenedge]

Hefty article about MRPs:-

[johnloony]

I have listed on my list 57 opinion polls since the election was called, some of which have several thousand in the sample rather than just one thousand. They are all ostensibly evenly spread across the whole country. From these, some people are trying to extrapolate (with MRP methods etc) to guess or,predict the results in individual constituencies.

It occurs to me that it would be probably just as accurate, and just as interesting, if the pollsters did it the other way round - doing constituency polls in individual constituencies, with a sample size of 1,000 each, and then trying to aggregate or extrapolate them to work out a national vote share equivalent. They would have to be careful in choosing a random selection of constituencies, including the boring safe seats, rather than concentrating on the marginals and interesting or unusual ones.

[jamesdoyle]

Jun 14, 2024 15:50:47 GMT hullenedge said:

Hefty article about MRPs:-

That could really have done with some editing. Some really tortured English, often saying something completely different to what he means; and there are references to diagrams which patently don't show what he says they show, I assume because he's tidied up his graphics and forgotten to amend the text to match.

[hullenedge]

Link to compare the MRPs:-

[Tony Otim]

Jun 20, 2024 14:02:09 GMT hullenedge said:

Link to compare the MRPs:-

Incredibly (or perhaps not), there are only 20 seats that stay Conservative in all models:

1. Arundel and South Downs
2. Beaconsfield
3. Brentwood and Ongar
4. Christchurch
5. East Grinstead & Uckfirld
6. East Hampshire
7. East Surrey
8. East Wiltshire
9. Epping Forest
10. Hamble Valley
11. Kingswinford & South Staffordshire
12. Maldon
13. Mid Buckinghamshire
14. New Forest East
15. North East Hampshire
16. North West Essex
17. Runnymede & Weybridge
18. Sevenoaks
19. South Shropshire
20. Weald of Kent

[hullenedge]

[jamesdoyle]

Jun 20, 2024 14:28:24 GMT Tony Otim said:

Jun 20, 2024 14:02:09 GMT hullenedge said:

Link to compare the MRPs:-

Incredibly (or perhaps not), there are only 20 seats that stay Conservative in all models:

1. Arundel and South Downs
2. Beaconsfield
3. Brentwood and Ongar
4. Christchurch
5. East Grinstead & Uckfirld
6. East Hampshire
7. East Surrey
8. East Wiltshire
9. Epping Forest
10. Hamble Valley
11. Kingswinford & South Staffordshire
12. Maldon
13. Mid Buckinghamshire
14. New Forest East
15. North East Hampshire
16. North West Essex
17. Runnymede & Weybridge
18. Sevenoaks
19. South Shropshire
20. Weald of Kent

Hmmm - 1 West, 2 North, 3 South, 6 East. The East is Blue?

[finsobruce]

Jun 24, 2024 14:15:07 GMT jamesdoyle said:

Jun 20, 2024 14:28:24 GMT Tony Otim said:

Incredibly (or perhaps not), there are only 20 seats that stay Conservative in all models:

1. Arundel and South Downs
2. Beaconsfield
3. Brentwood and Ongar
4. Christchurch
5. East Grinstead & Uckfirld
6. East Hampshire
7. East Surrey
8. East Wiltshire
9. Epping Forest
10. Hamble Valley
11. Kingswinford & South Staffordshire
12. Maldon
13. Mid Buckinghamshire
14. New Forest East
15. North East Hampshire
16. North West Essex
17. Runnymede & Weybridge
18. Sevenoaks
19. South Shropshire
20. Weald of Kent

Hmmm - 1 West, 2 North, 3 South, 6 East. The East is Blue?

Oh, East is East, and West is West, and never the twain shall meet,
Till Earth and Sky stand presently where the Tories win a Seat;
But there is neither East nor West, Border, nor Breed, nor Birth,

When two strong candidates stand face to face to find an easy berth! *

[hullenedge]

Another poll, with a different outcome, of Jewish voters:-

[The Bishop]

Well, both polls agree that Labour has markedly improved its position with them compared to the 2019 GE.

I wonder if the definition of "Jewish" was the same in each?

[batman]

It has to be, but it’s not unknown for people to say they’re Jewish when they are not halachically so