Polls showed 33% for labour and 34% for the Conservatives - as it was they got 30% and 37% with the Labour overstated by 3% and the Conservatives understated by 3% and the error rate on polls 3%
I know what you are saying but the fact that all the polls converged on the 33-33 or 34-33 figure shows that somewhere something went wrong, if the polls had of been successful you'd of expected it to be more like 36-32 or 37-31 with just the odd poll showing it on the margin of error boundaries.
Also there were a small number of polls through the election period that produced a 6-7% Con lead - yet some of these were suppressed from publication, and others were sent out with a load of caveats as to why they might be wrong. They might not have been 'right' for the 'right' reason but poll companies seemed not to want to be too different from one another.
Also there were a small number of polls through the election period that produced a 6-7% Con lead - yet some of these were suppressed from publication, and others were sent out with a load of caveats as to why they might be wrong. They might not have been 'right' for the 'right' reason but poll companies seemed not to want to be too different from one another.
I agree it does seem they "herded" together in fear of being wrong towards the back end of the campaign.
Polls showed 33% for labour and 34% for the Conservatives - as it was they got 30% and 37% with the Labour overstated by 3% and the Conservatives understated by 3% and the error rate on polls 3%
Most of the polls were wrong and outside the margin of error for the Conservative's share. A lot of people make the mistake of using the UK wide share but the correct actual figure to use when comparing the performance of polls is the GB share as polls don't include NI so it should be a rounded Con 38, Lab 31.
For the record, the GB wide actual shares to one decimal place which should be used to check the polls are:
Secondly, as stated above, whilst a 3% margin of error applies to any particular poll, this should be random if it's genuine sampling error rather than systematic bias. Given how many polls we had, if they were correctly predicting the final shares then they should have been scattered around the correct figures i.e. we would have had a similar amount of polls showing Con above 38 as below.
Of the final polls by eleven different pollsters, the average share was 33.5 for both Con and Lab which is way out for Con. Even if we take the polls individually, nine out of the eleven had either Con or Lab outside the 3% MOE (eight of them had Con outside the MOE). Only two managed to get both Con and Lab within the 3% MOE (ComRes and Opinium both had Con 35, Lab 34) and even then it was only within the extreme limits of the MOE.
Also takes a swipe at the political commentators. Mr Crosby accused political commentators of just wanting “entertainment” from politics and said that the “last time they met a punter was when they picked up their dry cleaning”.
“They were tested and found wanting,” Mr Crosby said.
First YouGov/Sun GE20 voting poll after the election
Con 41 Lab 30 LD 7 UKIP 13 GRN 4
Call a snap election Dave....
Does anyone know if YouGov have changed their methodology? Does this poll take into account the "shy Tory" factor? Because if it doesn't the scary thought for Labour supporters is the reality could be even worse.
However in as much as the credibility levels of the polls allow (very low at the moment) these figures tell me
A) There has not been the "recoil in horror" reaction to the Queens speech that many forum members were predicting.
Where do the "Labour must move left" advocates think their support will come from? The Green vote maybe? Well even if every single Green voter switches that would give them 34%. Can the LibDem vote be trimmed further? UKIP? Surely a move left would mean more multiculturalism and looser immigration controls? Is this likely to attract back ex-Labour kippers? Is it just to win back Scotland? Well if, as some are suggesting, the Tory victory was because of English voters fear of the SNP, how will appeasing Scotand play out in England, which has 10 times the available seats?
Does anyone know if YouGov have changed their methodology? Does this poll take into account the "shy Tory" factor? Because if it doesn't the scary thought for Labour supporters is the reality could be even worse.
However in as much as the credibility levels of the polls allow (very low at the moment) these figures tell me
A) There has not been the "recoil in horror" reaction to the Queens speech that many forum members were predicting.
Where do the "Labour must move left" advocates think their support will come from? The Green vote maybe? Well even if every single Green voter switches that would give them 34%. Can the LibDem vote be trimmed further? UKIP? Surely a move left would mean more multiculturalism and looser immigration controls? Is this likely to attract back ex-Labour kippers? Is it just to win back Scotland? Well if, as some are suggesting, the Tory victory was because of English voters fear of the SNP, how will appeasing Scotand play out in England, which has 10 times the available seats?
Really really worrying times for Labour I think.
Polls are now weighed to the 2015 election instead of 2010
Comments
I know what you are saying but the fact that all the polls converged on the 33-33 or 34-33 figure shows that somewhere something went wrong, if the polls had of been successful you'd of expected it to be more like 36-32 or 37-31 with just the odd poll showing it on the margin of error boundaries.
I agree it does seem they "herded" together in fear of being wrong towards the back end of the campaign.
Most of the polls were wrong and outside the margin of error for the Conservative's share. A lot of people make the mistake of using the UK wide share but the correct actual figure to use when comparing the performance of polls is the GB share as polls don't include NI so it should be a rounded Con 38, Lab 31.
For the record, the GB wide actual shares to one decimal place which should be used to check the polls are:
Con - 37.8
Lab - 31.2
LD - 8.1
UKIP - 12.9
GRN - 3.8
Secondly, as stated above, whilst a 3% margin of error applies to any particular poll, this should be random if it's genuine sampling error rather than systematic bias. Given how many polls we had, if they were correctly predicting the final shares then they should have been scattered around the correct figures i.e. we would have had a similar amount of polls showing Con above 38 as below.
Of the final polls by eleven different pollsters, the average share was 33.5 for both Con and Lab which is way out for Con. Even if we take the polls individually, nine out of the eleven had either Con or Lab outside the 3% MOE (eight of them had Con outside the MOE). Only two managed to get both Con and Lab within the 3% MOE (ComRes and Opinium both had Con 35, Lab 34) and even then it was only within the extreme limits of the MOE.
So yes, the pollsters got it wrong.
http://blogs.spectator.co.uk/robertsmith/2015/05/lord-ashcrofts-polling-overlooked-many-of-the-real-election-battlegrounds/
http://www.telegraph.co.uk/news/general-election-2015/11608648/Lynton-Crosby-Betrayal-of-British-voters.html
Also takes a swipe at the political commentators.
Mr Crosby accused political commentators of just wanting “entertainment” from politics and said that the “last time they met a punter was when they picked up their dry cleaning”.
“They were tested and found wanting,” Mr Crosby said.
Anyone want to have a guess about the relative state of the parties?
I would think that the Tories have gone up a point or two, labour static, the Libdems up and UKIP and the greens down.
Is there nothing this lot don't wish to ban?!
Total and utter control freaks. Scary.
If they were banned then you would no longer need to go into hiding on Election Night.
The party you support are a bunch of authoritarian control freaks.
So being locked up for 90 days = ok
random browser sniffing - not ok
?
Well......the party you support are currently busy dumping all the policies that featured in their Election Manifesto.
The party you support are a bunch of authoritarian control freaks
All about the referendum polling wise now
First YouGov/Sun GE20 voting poll after the election
Con 41 Lab 30 LD 7 UKIP 13 GRN 4
Call a snap election Dave....
It's wrong. lol.
YouGov owe it to us to say whether they've fiddled with their algorithms (ooh Matron etc.)
Except Boris wouldn't!
Has Jol read this yet ?:D:D:D Ho! Ho! Ho! Labour are stuffed
http://labour-uncut.co.uk/2015/05/02/revealed-eds-night-time-dash-to-casa-brand-driven-by-postal-ballot-panic/
There is even talk of seats being lower than in 2010 which is exactly what happened.
Opinion Polls what the hell was going on at You Gov et al?
Does anyone know if YouGov have changed their methodology? Does this poll take into account the "shy Tory" factor? Because if it doesn't the scary thought for Labour supporters is the reality could be even worse.
However in as much as the credibility levels of the polls allow (very low at the moment) these figures tell me
A) There has not been the "recoil in horror" reaction to the Queens speech that many forum members were predicting.
Really really worrying times for Labour I think.
Polls are now weighed to the 2015 election instead of 2010
Lovely!:kitty::kitty::kitty: