redsquirrel
This Machine Kills Progressives
At the height of the Tory peak, some of those seats in the NE looked, if not vulnerable, possible - Hartlepool for instance.
Political polling
- Thread starter brogdale
- Start date
The ICM polls have been outliers from the beginning, giving the Tories a much bigger lead than other polls. As others have said, it is the direction of travel which matters - based on recent polls however, it does seem that Labour's support may have reached it's peak, or maybe Corbyn's terrorism speech hasn't gone down well.
bi0boy
Power User
Because the polls are so methodologically flawed, it generally turns out that one of the 'outliers' is closest to the actual result. Which side of the polling average that will be this time round is of course a matter of speculation.
free spirit
more tea vicar?
Some further analysis on the rise of the youth vote...
In the IPSO mori polls, comparing the latest poll with the last poll before the 2015 election...
In the under 25s category there's been a jump to 90% saying they're between 8-10 our of 10 like ly to vote vs 77% in 2015.
In the 25-35 category that's now at 87% between 8-10 out of 10 likely to vote vs 83% last time.
Interestingly the 35-44 age range likelihood to vote has fallen from 81% to 73%, and the upper age ranges have also fallen slightly (though may not be statistically significant)
One factor that is going to skew the polls is that 29% of 18-24s have picked 8 out of 10 as their likelihood to vote, which is a far higher level than any other age range (no others are in double figures), but the headline voting figures for Ipso come only from those who answered 9 or 10/10 likely to vote.
https://www.ipsos.com/sites/default...ssets/Docs/Polls/polmon_april15_vi_tables.pdf
https://www.ipsos.com/sites/default/files/2017-05/pm-may-2017-tables.pdf
In the IPSO mori polls, comparing the latest poll with the last poll before the 2015 election...
In the under 25s category there's been a jump to 90% saying they're between 8-10 our of 10 like ly to vote vs 77% in 2015.
In the 25-35 category that's now at 87% between 8-10 out of 10 likely to vote vs 83% last time.
Interestingly the 35-44 age range likelihood to vote has fallen from 81% to 73%, and the upper age ranges have also fallen slightly (though may not be statistically significant)
One factor that is going to skew the polls is that 29% of 18-24s have picked 8 out of 10 as their likelihood to vote, which is a far higher level than any other age range (no others are in double figures), but the headline voting figures for Ipso come only from those who answered 9 or 10/10 likely to vote.
https://www.ipsos.com/sites/default...ssets/Docs/Polls/polmon_april15_vi_tables.pdf
https://www.ipsos.com/sites/default/files/2017-05/pm-may-2017-tables.pdf
Kaka Tim
Half Arsed and Slapdash till I Die
yeah - prior to manchester Teresa may was on the ropes and labour had the momentum. My feeling is that the polls wont change much now.
Kaka Tim
Half Arsed and Slapdash till I Die
Latest poll from Survation
CON: 43% (-)
LAB: 37% (+3)
LDEM: 8% (-)
UKIP: 4% (-) (26-27 May)
looks like labour is gaining from the "dont knows" rather than the tories - and probably cant squeeze much more out of the lib dems/UKIP/greens.
All We need is a UKIP revival to take 5% of the tories and its the red flag over downing street!
CON: 43% (-)
LAB: 37% (+3)
LDEM: 8% (-)
UKIP: 4% (-) (26-27 May)
looks like labour is gaining from the "dont knows" rather than the tories - and probably cant squeeze much more out of the lib dems/UKIP/greens.
All We need is a UKIP revival to take 5% of the tories and its the red flag over downing street!
free spirit
more tea vicar?
What we need is more of the former UKIP vote to end up balking at the thought of voting Tory when it comes to it and either staying home, switching back to Labour or voting UKIP, but there/s no option for voting UKIP in 200 of the key seats so it's mainly down to the first 2 options.
Though some of the polling is showing a bit of a swing from UKIP to Green, presumably as an anti-establishment vote.
Looks like the Greens are going to take a battering in most places, but hopefully can do enough on a local level in a couple of places to end up with an extra MP or 2.
Though some of the polling is showing a bit of a swing from UKIP to Green, presumably as an anti-establishment vote.
Looks like the Greens are going to take a battering in most places, but hopefully can do enough on a local level in a couple of places to end up with an extra MP or 2.
UKIP are offering a national Dementia Strategy/Service, as well as some other policies that may seem attractive to certain cohorts.
I do wonder whether some of those extra % points for the Tories are very soft indeed and won't turn out for them - last minute stay at homes or switching to older habits in the booth.
Really can't see the Greens winning Bristol West despite them trying to hype it. West feels like an increased Labour vote going by friends there and looking at the poster game - this is one constituency where people aren't shy about putting up posters and stakeboards for their choices. The students will seal it for Labour.
I do also wonder whether there's some kind of 'deny May a landslide' effect that could end up denying them even more than that - people who aren't wildly opposed to the Tories but don't want them to have carte blanche. If enough people in the right places think this way then they lose their majority entirely. Just thinking out loud really
mikey mikey
Active Member
An attempt to get the Tory vote out using fear?
free spirit
more tea vicar?
unfortunately it works the other way with labour voters who've been convinced that corbyn can't win, so there's no point voting.
My dad is lifelong Tory but is long past his auto-Toryism. Has voted Lib Dem in the past but is seriously considering a Labour vote in order to unseat Charlotte Leslie over Europe amongst other things. He seems very unimpressed by May & co and has warmed to JC as the campaign has gone on - at least seeing the irrelevance of decades old actions and the transparency of the campaign against him.
I do think newbie has a point though! It's very hard to vote for a hung parliament or small majority under fptp and a couple of % points can swing the result into solid win/landslide territory.
newbie
undisambiguated
I'm desperate to keep the bearded terrorist out of No10
-v-
I'm desperate for a Labour government but can't be arsed to vote
Dunno but it stinks of expectation management and trying to firm up the wobbly periphery of the Tory vote
Vintage Paw
dead stare and computer glare
Polling/voting of 40% Tory and 41% Labour won't put Labour in power. Thanks to FPTP and Scotland, Labour need a considerable swing and to be several points ahead of the Tories to get a majority of 1.
Vintage Paw
dead stare and computer glare
free spirit
more tea vicar?
indeed, which is why Labour really need to get their heads around the idea of how to make a minority Labour government work with support from SNP, Plaid, Greens and erm well maybe not lib dems who seem intent on sitting on the fence no matter what
Raheem
Well-Known Member
Actually, 41%-40% could just about do it, I reckon. It would probably take a little more if the swing is uniform across the country, but you would only need the distribution to be slightly favourable to Labour to make it work.
Assuming that this is correct. The polling companies were underestimated the Labour vote and overestimated the Conservative vote in 2015 because of unrepresentative samples - they interviewed too many Labour supporters and not enough Conservative supporters. Shy Tories (who don't want to admit they vote Tory) and lazy Labour supporters (who say they will vote Labour but don't) appear to be myths.
My thinking is that the same problems exist now as they did two years ago so the polls will still be overestimating the Labour vote and underestimating the Conservative vote. The problem is now compounded with polling companies adding in various fudge factors to counter the Labour bias ie. a lot of tweeking of turnout models.
At present I trust Survation the most because their data is the least processed (looking at it as outsider looking in). With YouGov, MORI and Panelbase being the next best bets (but who knows how to assess these fudge factors?). However remember that Survation are likely to be overestimating Labour and underestimating the Tories because of the underlying problems of polling. So take a couple of percent off Labour's Survation polling and add it to the Tories. That's my rule of thumb. It's a fudge and a crude fudge at that, but at least it's my own and I know about it.
Anyway if I was in charge of a polling company this is more or less how I would do it:
http://survation.com/methodology/
Ie. Weight your sample to make it match the population by all means but don't mess with the turnout model which could well be more fluid than you think. Go by what people tell you.
Take that as a recommendation or a disrecommendation as you please.
My thinking is that the same problems exist now as they did two years ago so the polls will still be overestimating the Labour vote and underestimating the Conservative vote. The problem is now compounded with polling companies adding in various fudge factors to counter the Labour bias ie. a lot of tweeking of turnout models.
At present I trust Survation the most because their data is the least processed (looking at it as outsider looking in). With YouGov, MORI and Panelbase being the next best bets (but who knows how to assess these fudge factors?). However remember that Survation are likely to be overestimating Labour and underestimating the Tories because of the underlying problems of polling. So take a couple of percent off Labour's Survation polling and add it to the Tories. That's my rule of thumb. It's a fudge and a crude fudge at that, but at least it's my own and I know about it.
Anyway if I was in charge of a polling company this is more or less how I would do it:
http://survation.com/methodology/
Ie. Weight your sample to make it match the population by all means but don't mess with the turnout model which could well be more fluid than you think. Go by what people tell you.
Take that as a recommendation or a disrecommendation as you please.
Vintage Paw
dead stare and computer glare
It's because of FTPT and the fact that distribution matters that means a straight up % majority isn't enough to do it. To win a majority of just one, Labour needs a swing akin to that of Blair's in 1997. That's not going to happen unless something remarkable is around the corner.
Here's a thing on regional issues: The polls could be right about Labour's gains - but also misleading
For some perspective, here's a list of Labour targets and the numbers involved (although it doesn't show vote share of other parties, so any UKIP votes going to the Tories are still going to muddy the waters): Labour Target Seats 2017 - Election Polling
I reiterate my recommendation of following NumbrCrunchrPolitics (@NCPoliticsUK) on Twitter - they've got lots of useful data.
free spirit
more tea vicar?
one thing's seeming increasingly clear - turnout will be far higher than at any recent election.
The latest Survation poll (which puts labour 6% behind) now has 48.5% of people who didn't vote in 2015 saying they're 10/10 likely to vote this time.
They also put 64% of those who're currently undecided as being 10/10 definitely going to vote.
Labours vote is also solidifying, with 89.1% now 10/10 likely to vote vs 87.9% of the tory vote, whereas in the last poll it was only 79% and lower than the tories.
The latest Survation poll (which puts labour 6% behind) now has 48.5% of people who didn't vote in 2015 saying they're 10/10 likely to vote this time.
They also put 64% of those who're currently undecided as being 10/10 definitely going to vote.
Labours vote is also solidifying, with 89.1% now 10/10 likely to vote vs 87.9% of the tory vote, whereas in the last poll it was only 79% and lower than the tories.
bi0boy
Power User
That just shows how much people lie or how unrepresentative the sample is. The 10/10 certainty to vote numbers imply turnout will be greater than 80%, which it won't be.
I can't find information about what percentage of eligible voters are registered this time, but there are reports that voter registration is up significantly and I don't think 80% turnout would be particularly surprising. 2015 was 66%, and this time registration is higher because of the EU referendum and because of a huge registration drive amongst young people, who are far more motivated this time as there is a meaningful difference between the parties. Turnout in this election will be significantly higher than in 2015, not just by a couple of percentage points. It may not be as high as 80%, but I would be surprised if it is less than 75%.
Similar threads
