Categories
politicalbetting.com » Blog Archive » Why this could be like 1992 when the polls were simply wron
politicalbetting.com » Blog Archive » Why this could be like 1992 when the polls were simply wrong?
Elections & polling expert Keiran Pedley examines whether the UK polling industry could be about to experience a crisis not seen since the polls got it so wrong in 1992.
0
This discussion has been closed.
Comments
On turnout - the fact that so many UKIPers did not vote last time strikes me as being pretty important; but that's less the case with SNP supporters as they will have voted in the referendum. I'd have thought that if anyone does not get out to vote in Scotland this time it will be Labour supporters in Labour heartlands. They must be totally demoralised.
Meanwhile, the shy Tory phenomenon could be most pronounced in the Midlands. Nothing in sub-samples (I know) or the local or EU elections indicates any groundswell of support for Labour.
As I keep saying, Labour will do very well indeed to stand still in this election. I expect them to go backwards.
Looking at the parties' official Youtube channels, it does seem to me that the Conservatives have done a better job than Labour, including with what Americans would call its "attack ads". This might be important if the party can get 50/50 voters to look at them -- and presumably this is where its American consultants' skills in identifying and targeting these voters comes into play. If it can, then late converts might better explain "shy Tories".
I do agree the additional variables added to this election make it harder to predict, so the polling may not be as accurate as in the past election. I'm especially interested to see how accurate the exit polls get it this time. I'm expecting a less perfect call this time, but the polls will be nowhere near far enough out to give any one party a working majority. Anything else is just wishful thinking.
Almost all the different polling companies, with their different methods and means of collecting voter intentions are closely aligned. Like last election, we get the occasional poll that goes way off in a different direction - but soon afterwards there's a correction back to mean of all pollsters. It's pretty clear where this election is headed.
I guess people will forever go on wanting to doubt the polls. We've seen a lot of denial through the Blair years, the last election here, and the last two in the USA. One such doubter in the U.S. predicted a landslide for Romney days before the last election, but ultimately polls and the experts called it almost perfectly; Dick Morris was left licking his wounds for making such a ludicrous call.
Common sense says there are mind-bogglingly more variables in 2015 than at probably any other election in history. Scope for error must, on the face of it, increase.
Agree with Mike about shy LibDems. Could be the biggest surprise of the night...
"UKIP is picking up support from those who distrust politicians, but crucially, this support is significantly higher among those distrusting of MPs and who tend to vote in general elections. The lesson we can take away from the BES internet panel is therefore that UKIP’s support is coming from those people who are politically engaged but disillusioned, much more than from those people who are politically disengaged and disillusioned."
http://www.britishelectionstudy.com/bes-resources/is-nigel-farage-the-heineken-politician-is-ukip-reaching-the-parts-of-the-electorate-other-parties-cannot-reach/
It seems as though you want to eat your cake and still have it with that post.
http://en.wikipedia.org/wiki/Eastleigh_by-election,_2013#Polling
I take it your guessing that the party will gain fewer than this number of seats in England & Wales and will therefore produce a negative outcome in terms of their overall UK seats tally compared wth 2010.
The best current odds come from Ladbrokes who offer 6/1 for Labour to win between 226 - 250 seats. Those preferring the next 25 seats band up, covering between 251-275 seats are best served by Hills' 2/1. Taken together, a combination bet covering both these bands would return odds of 11/10.
I start Monday reading the best oxymoron of the month so far :-)
Donkeys years of experience of trying to get these people to engage suggests to me that, when you do finally get them to lower their defences a little and show their true allegiance, they are disproportionately Conservatives. Although I am prepared to admit, this time around, they could be hiding a fair few Kippers too.
And this time around, in my experience from canvassing an ultra-tight LibDem marginal, the LibDems aren't shy. They are dispirited. Broken. Can't see the point any more. In many cases have been lost to the Can't Be Arsed Party. The LibDem challenge is to get them to vote at all.
The recent SW England ComRes poll seems to agree with the Ashcroft 1st question result.
Be curious to see what polls undertaken on their previous basis would be showing. If it were significantly different, would they be big enough to admit so - and revert?
The ones I can immediately think of are UK 1992 and Israel 2015, both of which were:
1) More conservative than polled.
2) Less change than the polls thought.
These would argue for both Con in England and Lab in Scotland out-performing, but British polling may have been innoculated against the former failure mode by 1992.
What other examples can people think of?
Titter ....
This would be most marked in the Con-Lab shares compared with the multiplicity of smaller parties. The directions of error in these are less likely to cancel out.
I'm off within the hour for a "briefing breakfast" (horrible term) with a very well placed source from national team Con.
I shall place these blue musings before the PB masses later this morning.
You lucky, lucky people.
This means that the polls are more likely to be wrong for this election even if the pollsters have tackled the issues in the article well. Great article, btw.
Phone polling has its own problems and it may be that the massive increase in caller ID since the last election will turn out to be a significant factor. Heaven help anyone trying to phone us on an international number these days.
The risk is that someone gets lucky and gets very close to the actual result with the result their methodology becoming the gold standard going forward when it was just a fluke. If this happens things will not get any better the next time. Given all their problems it seems to me to be seriously impressive that the polling companies get as close as they do. There are a lot of very clever people in these companies doing their absolute best to get things right. I see none of the systematic bias one sees in the US where so many pollsters seem to consistently favour one party or the other.
FWIW although it is possibly overdone I suspect that ICM has been successful for so long because they are right. Did not vote do not vote and any party (even the SNP) heavily reliant on such support is riding for a fall.
A despised politician, given his jotters by the electorate, gets to keep his grasping mitts on power thanks to the corrupt system of appointed Lords. A man the voters decided was unfit for office gets to continue to make laws and rule over the people who rejected him.
It is utterly pathetic.
It also makes calling turnout difficult. I am sure that it will be low in absolute numbers, but if a lot of the absentees are unregistered show as a high percentage. I still favour the 60-65% range though.
That’s not to say I don’t think that the whole system needs a complete change.
I've always said that these polls are wrong - for whatever reason - and it now seems that my fears will be proven right.
Kippers are blessed with an evangelical zeal.
My guess would be that the fall in turnout will disproportionately hit the SNP as it will largely come from those whose vote in the referendum was a one shot deal. This will only affect the margin of their victories in most cases of course.
What the increased registration does mean is that if the next set of boundaries gets created on the current (post-Tory-reform) rules, Scotland will be massively over-represented compared to its population.
It's quite an impressive feat of political cunning that Lab, Lib-Con have each had a turn at messing around with the boundaries in the hope of benefitting themselves, and somehow managed to create a huge win for the SNP at their own expense.
Shy Libs, Kippers, Tories -- surely the number of billboards and posters outside peoples homes gives an indication of this?
I'd say there were more Tory billboards this election than I have ever seen before, and less 'Winning Heres' in this election. So, I am more inclined to believe in Shy LibDems than Tories in this election.
"However, if the industry does experience another ‘Shy Tory’ 1992 moment then expect to hear a lot more from Lord Foulkes in the coming weeks and months and his plans to introduce a statutory body to regular the polling industry. Then things really will get interesting."
Just when I had forgotten about one of the biggest morons in New Labour, up pops Lord Foulkes.
The danger for the Tories is this last minute shift could now be going directly to UKIP as the recipient of the right.
Was there not, for a period of 40 or so years ago, in the late 19th & early 20th Centuries, a very significant block of Irish Nationalists in the Commons? And governments of both main parties had to take their views into account when framing legislation!
1) Some of the pollsters are definitely going to be wrong as things stand, because the disagreement about how UKIP is performing is enormous, with three weeks to go.
2) Conversely, the sheer consistency of the polling in Scotland makes it less likely that the polling is completely wrong there.
A lot to ponder upon.
I'm going to bookmark that link for the next time a politician says the world is going to end because the kids are addicted to the new thing.
It's never the disaster they think it will be.
The point before about potential UKIP supporters being unwilling to admit to a phone pollster that they would vote UKIP on embarrassment grounds is a valid one, but I think is more common among "red" northern WWC UKIP types than "blue" southern ones: having grown up in the former type of household, it is hard to explain to people on the outside just how much folk are worried about what people will think of them if they give the "wrong" answer. It may explain why UKIP support tends to go down when there is a furore about UKIP; it might also explain why Ashcroft got Clacton right ("Blue" kippers) but got H&M totally wrong (shy "red" kippers).
One other thing: was campaigning in a marginal Conservative seat in the Midlands at the weekend. Feedback was that, while they are not taking anything for granted, they feel quietly confident. They are getting quite a bit of anti-Cameron feedback, but no one is energised by Labour so the feeling is a lot of these will not vote, which they are fine with...also said Sturgeon is being raised by a lot of people on the doorstep and is proving to be a significant factor in influencing votes. Only seat they were worried about in the area was Cannock Chase but, even there, the tone was more positive than would have expected i.e. they felt it could still be held whereas I would have thought it was a goner.
http://www.dailyrecord.co.uk/news/politics/fanatical-snp-councillor-slams-colleagues-5551628
This probably won't get traction because national trends seem to swamp local matters in Scotland, but the SNP in this area seems to be a nest of vipers.
Interesting article, and I was unaware there were moves afoot to regulate the polling industry. How would that change things? Would it involve a reduction in polling during a campaign?
For those who missed it, here's my post-race analysis of Bahrain:
http://enormo-haddock.blogspot.co.uk/2015/04/bahrain-post-race-analysis.html
The season's effectively in quarters, with a first fly-away section (now complete), the European middle cut in two by the mid-season interval, and a final few back-to-back races. Oddly, I've been red in every race but by such small margins (last time was 50p down, assuming a £10 stake) it's difficult to be either happy or overly displeased.
It's another baseless smear. And will have the same effect as the previous attempts. The DayLate Record is already a laughing stock and it's sub 200k readership matters less and less.
Two high profile Labour MPs that could be in danger... Just how much traction does the propaganda wing of the Scottish Labour party really have these days though ?
Can the Record influence floating voters - that is the real question.
The Tories will not be polling more than 36%. The best they can hope for is having the most seats by a much smaller margin than last time.
The LibDems won't be making them Kings this time either.
I disagree with you. It's plausible (although unlikely) that the Conservatives could maintain or increase their number of seats. It's also possible they could lost some but have a larger margin over Labour [although that's even less likely].
ICM had them 36-36-39 in the last three polls, so 36% plus is not impossible.
What I find disturbing is the element of reward for failure - not a good management strategy.
The survey of more than 13,000 final-year students at 30 universities — conducted by recruitment analysts High Fliers Research — found that the Greens were the next most popular party, with 25 per cent support, followed by just 6 per cent for the Liberal Democrats and 1 per cent for the UK Independence party.
This represents a significant drop in Lib Dem popularity compared with the same survey in 2010, which found that 23 per cent of students were intending to vote for Nick Clegg’s party.
Students have moved away from the Lib Dems after the party broke their 2010 election pledge not to increase tuition fees.
The poll also divided universities by their majority political leanings, showing that support for the Tories was strongest at Imperial College London, the London School of Economics, Durham, Bath, Exeter, and Loughborough — a key marginal seat held by the Tory education secretary Nicky Morgan.
Meanwhile, Oxford, Cambridge, Warwick, Liverpool, Lancaster, Manchester and Sheffield universities were all dominated by Labour voters.
http://on.ft.com/1yJSeiw
Scottish Sun:
FLUSTERED David Cameron launched a rant against the Nats on live TV yesterday as Nicola Sturgeon watched backstage.
The PM lost his cool as he told the BBC’s Andrew Marr Show that the prospect of the SNP propping up a Labour government was “frightening”.
I can only read the toplines as I'm not a Sun subber but this headline is meat and drink to SNP supporters I'd have thought.
Of course it all boils down to Paper sales - with the Record pumping out Murphy's propaganda and the Tories forever on 16% there is a gap in the market which the Sun has filled.
Shy SLAB and lower SNP turnout? It certainly seems possible simply because the SNP position is so amazingly great that it would not out of the realms of possibility there is a little fall back from that. On the one hand the change of only a few percent could save quite a few seats, but on the other, there was a lot of thought that the SNP would fall back from where they were some months ago for the same reasons, and it hasn't happened, in fact their position has gotten better.
Shy Tories? I'm not seeing the evidence for this, not to the point it should be automatically assumed it will happen, as seems to be the case with all these official Con plurality predictions.
Shy LDs? I used to believe in them, just like I believed their position would recover to the mid-low teens, enabling a GE day result of mid-late teens perhaps, if they were lucky. Instead their position has gotten even worse and, in my view, passed the point of no return. They will not be destroyed - though they will be deeply wounded - much as their opponents would like, they will still get circa 20 seats which though terrible is not nothing and something historically they have been around before an survived. But their position is so terrible in the polls that any 'shy' LDs, or any waverers, will probably see no point voting for them as even in what was once a safeish seat they appear likely to lose, and vote for someone else or stay home instead, making no seat seat (much like the SNP vs everyone - except funnily enough the Tories in a few seats - in Scotland position). I don't normally like to get so personally direct or rude, but he sounds like an absolute cretin.
FWIW (*doorstep anecdote alert*) I'm still, even at this fairly late stage, meeting vast numbers of undecided voters who previously voted Tory, saying things like "the Conservative Party has left me", or that they don't like a lot of what the party stands for.
Depending on if, and how late, these people decide to hold their collective noses and vote Conservative, this could easily look like a shy Tory effect.
Another coalition between the Conservatives and the Liberal Democrats is not going to happen, Andrew George, the Liberal Democrat MP for the highly marginal seat of St Ives has declared.
His leader, Nick Clegg, has insisted his party is determined to form a coalition with either the Tories or Labour. Clegg is due to campaign in the seat for George on Monday.
But George told a public meeting last week a second coalition with the Tories “is not going to happen. We have had enough of it. The Tories would not want it and I am sure my party would not go for it.”
http://bit.ly/1Hna8ZF
Enough people in Scotland do not read it that way to give the SNP 50% of the vote. If this sort of tactic worked, it would have worked already. Everyone knows the SNP still want independence, it's not a secret. Enough people also know they get decent, competent government as the SNP try to demonstrate the merits of independence.
While I agree with your analysis, it would be faintly amusing if the Tories got a little more than 36%, and would make me feel even more sorry for Cameron - the chances of him increasing his vote share from 2010 seemed very difficult given all the unpopular things governments generally have to do, particular during a time of austerity, and yet even if he achieves that, as you say the best he can hope for is most seats by a smaller margin than last time, and personally that looks unlikely to me too.
I'm also sure that whoever gets to be PM there is nothing - NOTHING - that will be in the SNP's interest to play nice. They don't come to Westminster to build bridges, they come to bury the UK. Being a good partner in the UK would destroy their USP. I am expecting an unending series of constitutional crises around Trident/Barnett/public spend per capita / whatever. The SNP should come to make itself and Scotland detested in England. That road leads to separation - which is, after all, precisely what they want.
A point to consider as the election nears is what sort of spin we might all be tempted to produce. There's the natural "we're gonna win!" spin, which just makes you feel good. There's the "it's very close, get out and vote" spin, which everyone uses for polling day (I remember someone's by-election leaflet for polling day afternoon which claimed that "early voting this morning shows it's very close", clearly printed days earlier - I refused to deliver this blatant lie). And there's even the "we're losing, and it'll be awful if you don't change your mind" spin, which seems to be the current Conservative strategy. The Nottingham Post asked me yesterday how we were doing, and I couldn't decide what I ought to spin, so in desperation I told him the truth (a guessed 5-6% lead). :-)
And I agree with him.
Some of those diamonds are f*cking enourmous mind.