johng
Labour
Posts: 4,850 
|
Post by johng on May 28, 2024 14:37:14 GMT
R&W promising a 'mega poll' with 12,000 respondents and regional breakdowns. To be released at 5pm.
|
|
nodealbrexiteer
Forum Regular
non aligned favour no deal brexit!
Posts: 4,456
|
Post by nodealbrexiteer on May 28, 2024 14:43:53 GMT
R&W promising a 'mega poll' with 12,000 respondents and regional breakdowns. To be released at 5pm. what a tease! |
|
cogload
Lib Dem
I jumped in the river and what did I see...
Posts: 9,144
|
Post by cogload on May 28, 2024 16:03:38 GMT
|
|
binky
Non-Aligned
Posts: 63
|
Post by binky on May 28, 2024 16:07:12 GMT
Where are the regional breakdowns?
|
|
|
|
Post by batman on May 28, 2024 16:07:12 GMT
very similar to the Survation poll but with a larger sample size. It's not surprising that the Tories are still losing ground because they are not giving most uncommitted voters any good reason to vote for them.
|
|
johng
Labour
Posts: 4,850
|
Post by johng on May 28, 2024 16:16:16 GMT
From their site: No regional breakdowns yet from what I can see.
|
|
|
|
Post by John Chanin on May 28, 2024 16:26:16 GMT
|
|
iain
Lib Dem
Posts: 11,452
|
Post by iain on May 28, 2024 16:57:28 GMT
From their site: No regional breakdowns yet from what I can see.
There are regional breakdowns if you click on the data tables. To be honest, they look little better than the usual random variation you get on crosstabs. |
|
johng
Labour
Posts: 4,850
|
Post by johng on May 28, 2024 17:20:12 GMT
There are regional breakdowns if you click on the data tables. To be honest, they look little better than the usual random variation you get on crosstabs.
Thanks! Not sure if I just missed it or it wasn't there.
Yeah, the regional samples are weighted - though some are quite small.
|
|
|
|
Post by greenrichard on May 28, 2024 17:25:44 GMT
I wish polling companies would realise we can handle numbers to a least two decimal places. A present it just looks wrong that we can have combined increase of 5pts and a combined decrease of 1pt.
|
|
|
|
Post by woollyliberal on May 28, 2024 18:05:02 GMT
I wish polling companies would realise we can handle numbers to a least two decimal places. A present it just looks wrong that we can have combined increase of 5pts and a combined decrease of 1pt. When the error could be +/- 3%, or 2% on very large sample polls, why is the difference between 6.0% and 6.2% of much use? |
|
|
|
Post by johnloony on May 28, 2024 19:38:35 GMT
I wish polling companies would realise we can handle numbers to a least two decimal places. A present it just looks wrong that we can have combined increase of 5pts and a combined decrease of 1pt. That’s not mathematically possible - even with a sample size of 12,000. |
|
|
|
Post by greenrichard on May 28, 2024 22:23:14 GMT
I wish polling companies would realise we can handle numbers to a least two decimal places. A present it just looks wrong that we can have combined increase of 5pts and a combined decrease of 1pt. That’s not mathematically possible - even with a sample size of 12,000. Which bit isn’t possible? Decimals or the fact this poll indicts that the combined rise was 5pts and the combined fall was 1%? |
|
|
|
Post by johnloony on May 29, 2024 0:37:53 GMT
That’s not mathematically possible - even with a sample size of 12,000. Which bit isn’t possible? Decimals or the fact this poll indicts that the combined rise was 5pts and the combined fall was 1%? It’s not possible to get to the level of accuracy which would be implied by decimals, if one were to quote the figures to an accuracy of one decimal place. The 95% confidence interval from even a properly randomly selected sample of 12,000 would be 0.9%. |
|
|
|
Post by johnloony on May 29, 2024 10:47:19 GMT
Regional totals, to the nearest whole %
n.b. The tables in the spreadsheet give separate figures for the Workers Party of Britain and "Others", but the total number of people saying they will vote for the WPB is only 48 out of 12,000 and in all regions it is less than 1%, so I have combined WPB and Others together.
GB Total--------: Con 23 Lab 46 Ref 13 LD 09 Grn 5
London----------: Con 26 Lab 45 Ref 10 LD 11 Grn 7 Oth 2
South East------: Con 26 Lab 40 Ref 13 LD 13 Grn 6 Oth 2
South West-----: Con 23 Lab 40 Ref 11 LD 20 Grn 4 Oth 2
Eastern----------: Con 22 Lab 46 Ref 17 LD 09 Grn 6 Oth 1
East Midlands-: Con 24 Lab 44 Ref 14 LD 13 Grn 4 Oth 1
West Midlands: Con 22 Lab 51 Ref 14 LD 08 Grn 4 Oth 1
Yorks & Humber: Con 21 Lab 51 Ref 18 LD 03 Grn 4 Oth 2
North West------: Con 17 Lab 58 Ref 11 LD 05 Grn 6 Oth 2
North East-------: Con 24 Lab 53 Ref 14 LD 06 Grn 3 Oth 1
Wales-------------: Con 25 Lab 45 Ref 13 LD 04 Grn 3 Oth 0 PC 10
Scotland----------: Con 22 Lab 35 Ref 07 LD 05 Grn 2 Oth 1 SNP 28
|
|
|
|
Post by johnloony on May 29, 2024 11:24:44 GMT
That means that the net changes for each party, compared with the actual result of the 2019 general election, are (to the nearest whole number):
GB Total--------: Con -22 Lab +13 Ref +11 LD -3 Grn +2
London----------: Con -06 Lab -03 Ref +09 LD -4 Grn +4
South East------: Con -28 Lab +18 Ref +13 LD -5 Grn +2
South West-----: Con -30 Lab +17 Ref +11 LD +2 Grn +0
Eastern----------: Con -35 Lab +22 Ref +17 LD -4 Grn +3
East Midlands-: Con -31 Lab +12 Ref +12 LD +5 Grn +1
West Midlands: Con -31 Lab +17 Ref +13 LD +0 Grn +1
Yorks & Humber: Con -22 Lab +12 Ref +12 LD -5 Grn +2
North West------: Con -20 Lab +11 Ref +07 LD -3 Grn +4
North East-------: Con -14 Lab +10 Ref +06 LD -1 Grn +1
Wales-------------: Con -11 Lab +4 Ref +08 LD -2 Grn +2 PC +0
Scotland----------: Con -3 Lab +16 Ref +06 LD -4 Grn +1 SNP -17
|
|
|
|
Post by johnloony on May 29, 2024 11:40:56 GMT
That means that the net changes for each party, compared with the actual result of the 2019 general election, are (to the nearest whole number): GB Total--------: Con -22 Lab +13 Ref +11 LD -3 Grn +2 London----------: Con -06 Lab -03 Ref +09 LD -4 Grn +4 South East------: Con -28 Lab +18 Ref +13 LD -5 Grn +2 South West-----: Con -30 Lab +17 Ref +11 LD +2 Grn +0 Eastern----------: Con -35 Lab +22 Ref +17 LD -4 Grn +3 East Midlands-: Con -31 Lab +12 Ref +12 LD +5 Grn +1 West Midlands: Con -31 Lab +17 Ref +13 LD +0 Grn +1 Yorks & Humber: Con -22 Lab +12 Ref +12 LD -5 Grn +2 North West------: Con -20 Lab +11 Ref +07 LD -3 Grn +4 North East-------: Con -14 Lab +10 Ref +06 LD -1 Grn +1 Wales-------------: Con -11 Lab +4 Ref +08 LD -2 Grn +2 PC +0 Scotland----------: Con -3 Lab +16 Ref +06 LD -4 Grn +1 SNP -17 Every cloud has a silver lining Every thunderstorm has a nugget of gold Every landslide has a slow patch If these figures are anything like accurate, it means that Sadiq Khan has already done so much damage to London that it will seriously damage Labour's chances of making advances in London constituencies like (frig zumple) Croydon South, which a uniform swing would put at risk. |
|
|
|
Post by batman on May 29, 2024 11:48:18 GMT
It would be extremely surprising if London became less Labour than the country as a whole as these regional samples suggest. I also find it extremely hard to credit that Labour has almost twice as many votes than the Tories in the SW, notwithstanding the latter's extreme weakness in much of Bristol.
|
|
|
|
Post by Pete Whitehead on May 29, 2024 12:27:29 GMT
Or that Labour are stronger and the Conservatives weaker in the East of England than in Wales. Clearly not weighted by region and not worth taking any notice of.
|
|
Crimson King
Lib Dem
Be nice to each other and sing in tune
Posts: 9,861
|
Post by Crimson King on May 29, 2024 13:42:49 GMT
Which bit isn’t possible? Decimals or the fact this poll indicts that the combined rise was 5pts and the combined fall was 1%? It’s not possible to get to the level of accuracy which would be implied by decimals, if one were to quote the figures to an accuracy of one decimal place. The 95% confidence interval from even a properly randomly selected sample of 12,000 would be 0.9%. I don’t think I agree. the decimal doesnt imply the accuracy of the prediction, it gives the result of the calculation, so a 95% ci from a figure of 23.3 would be 22.4 to 25.2 rounding gives 2 problems - firstly it gives the daft sutuation where 3 parties can be up “1%” each and one party down “1%” but also if you want to see the expected range of results you will be applying the ci to the wrong figure giving a range of 22.1 to 24.9 in the example above |
|
