D minus 25
Yesterday’s polls:
ICM/Sunday Telegraph (7th Apr) CON 38%(+1), LAB 30%(-3), LDEM 21%(nc)
YouGov/Sunday Times (9th-10th Apr) CON 40%(nc), LAB 32%(+2), LDEM 18%(-2)
ComRes/Independent on Sunday/Sunday Mirror (9th-10th Apr) CON 39%(+2), LAB 32%(+2), LDEM 16%(-4)
BPIX/Mail on Sunday (9th-10th Apr) CON 38%(+1), LAB 31%(+1), LDEM 20%(nc)

There was also a OnePoll survey in the People, which showed figures of CON 37%, LAB 31%, LDEM 21% but for which I don’t have the information to know if we can give it any weight.

The topline voting intention are all broadly consistent – largely I expect because yesterday’s polls all came from the more established pollsters and those showing lower levels of support for Labour have tended to be the new entrants to the business. The Conservatives are at or just below 40% (ranging between 38% and 40%), the Labour party in the low thirties (30% to 32%). There is more variation in the Lib Dem score, with ranges between 21% and 16% – that latter score just doesn’t ring true to me and I’d be surprised if ComRes’s next poll doesn’t show them bouncing back.

The Conservative lead of something around the 8 points that these polls imply (the equivalent of a 5.5% swing) would not be enough for an overall majority in itself. Rather the Conservatives would have to rely upon outperforming the national swing in the marginal seats that actually decide the election. We had one marginals poll last night, from ICM, and it showed a swing of 6.3% to the Conservatives, so only slightly larger than the national one and still slightly short of the 6.9% they need for an overall majority.

This morning’s Independent on Sunday has has predictions of the pollsters themselves. All except Ben Page of MORI predict a small overall Conservative majority (including Peter Kellner, Andrew Cooper, Martin Boon and Andrew Hawkins) – the implication being that the pollsters expect either the Tory lead to grow during the campaign, or the Conservatives to outperform in the marginals by more than yesterday’s ICM poll suggested.

UPDATE: In the comment below Ben Page of MORI has clarified that what the Indy had as his projection was actually what he thinks would happen based on the polls now. His prediction for the final result is also a small Conservative majority.


529 Responses to “Sunday morning round up”

1 5 6 7 8 9 11
  1. Rubbish ! One of them even suggested a 50 seat majority. Based on what ? You have to add to UNS a great chunk of wishful thinking from any recent poll to arrive at a 50 seat majority.”

    If Angus Reid’s 11% lead is the outcome on May 6th, I’d be surprised if a 40/50 seat majority was not the outcome.

    Peter Kellner is predicting 20/30 which isn’t far off that, either.

    Rob Sheffield – shy Labour voters may be very important. Without them, both Populus and ICM would give 10% leads for the Tories.

  2. Sue Marsh
    Sean Fear – “I imagine it’s because they identify trends in their polling, and have access to regional data, and data in specific types of seats, that isn’t necessarily published.
    At the end of the day, they’re measuring vote shares, not seat totals – but they have a great deal of knowledge of how vote shares are likely to translate into seat totals.”

    Sean Fear – thank you for a rational answer to what I saw as a strange occurrence without simply lambasting me with partisan nonsense. THAT was the answer I was looking for.
    —————————————————-
    Sue, here I cannot agree to you agreeing with Sean Fear.
    There are no rational grounds based on current data to extrapolate to a 50 seat majority as one of the pollster’s have predicted. To achieve that kind of swing would need tailor-made swings like Tories getting 3% swing when they needed 2% and so on and then hitting 9% swings when they needed 8.5%. Otherwise, it is mathematically not possible to achieve a 50 seat majority “on current published polls” from these same pollsters !!!

  3. If Cameron wants this election, one simple maneouvre would secure it. Labour have not included a pledge not to put up VAT.

    A gaping hole emerges right in the heart of their election strategy. If George Osborne can guarantee no VAT rise, he will be the next Chancellor.

    D’oh

  4. Surfdoc, an exit poll is a rather different animal from a normal opinion poll.

    An exit poll is asking people how they have voted, not how they will vote; it involves asking tens of thousands of people; and the aim is to predict seat totals accurately – vote share is secondary.

    The 2005 exit poll was outstandingly accurate.

  5. @EOIN – “If George Osborne can guarantee no VAT rise, he will be the next Chancellor.”
    People wouldn’t necessarily believe his guarantee though.

  6. What polls are being released tonight?

  7. Sean Fear

    If Angus Reid’s 11% lead is the outcome on May 6th, I’d be surprised if a 40/50 seat majority was not the outcome.
    —————————————
    My position stands. 11% lead is only equal to a 7% swing. The Tories need 6.9% swing just for a majority of one. They can only have a majority of 50 if they could achieve tailor-made majorities in most of the earlier 116 seats and use the “saved” swings in the next 25 seats. Some wishful thinking ! No wonder it is called ARSE. Even the polls come out of it !!

  8. Surbiton,

    In 1987, an 11% lead gave the Conservatives a 100 seat majority; in 1979, an 8% lead gave a majority of 43.

    So, to say that it’s not mathematically possible for AR’s 11% lead to produce a majority of 40/50 is demonstrably wrong.

  9. @Amber……………With your very endearing commitment to a crystal ball, and my dream polls, we might take polling debate to a whole new level. :-)

  10. Surbiton – I know, I’m not saying it makes me agree with the pollsters, simply that at least Sean Fear gives me a rational reason why it might be the case.

  11. Eoin Clarke
    “Still don’t know what a Quattro is”
    Eoin, Suzy Quatro is a great 70s Rock Star. OnePoll’s male voters say she has the nicest celeb bottom, just a head of Chris Grayling.

  12. sean fear

    ” shy Labour voters may be very important. Without them, both Populus and ICM would give 10% leads for the Tories”

    !!!!!!! Great stuff.

    Now you pipe in with some mystical unproven ‘assertion’ that of course completely misunderstands the whole concept: so let-me-explain-it-to you….

    ** ‘shy voters’ say one thing in a poll and do another thing in the polling booth** For example they say in a poll they will vote conservative and in the polling booth they vote for the government….

    So- of course- (if they exist) they are part of the current non-labour vote….and make for an overstatement of the Conservative support and lead in current OP’s.

    OK ?!

    You guys are really on fire today its been most enjoyable: now I am off to the pub to have a drink to Pompey…..and victory to all underdogs :-)

  13. Sean Fear
    Surbiton,

    In 1987, an 11% lead gave the Conservatives a 100 seat majority; in 1979, an 8% lead gave a majority of 43.

    So, to say that it’s not mathematically possible for AR’s 11% lead to produce a majority of 40/50 is demonstrably wrong.
    =========================
    Sorry, in 1987 they were not facing the same polling landscape. In 1951 Labour got 49% of the vote and higher than the Tories and still lost. Because the way the majorities were distributed were different. You cannot compare 1987 to 2010.

  14. To those who ask why, if the pollsters think their own polls are wrong, they don’t just alter them…

    Really? Are you serious? You think it would be sensisble or appropriate for Anthony to receive some new poll data and say “Hmm, that Labour figure looks a bit high to me, knock a couple of points off it”??

    I suspect what is at play here is a combination of three things.

    a) A hunch that the Tories may have a reasonably good campaign.

    b) A hunch that, as in some previous elections, the Tories may outperform the opinion polls at the Big Event.

    c) A hunch that the marginal effect is real and will add the equivalent of 1-2% to the Tory lead when it comes to seat calculations.

    I don’t necessarily share all of those views, but that’s my best guess as to what the pollsters are thinking when they predict a better Tory outcome than their own polls are currently showing. Of course they are just guessing really, but their guesses are amongst the most “educated” in country.

  15. Sean,

    You are undoing months of good posting with some dubious comments…..

    pull back from the brink (my advice)

    @Julian,

    if it is in the Tory manifesto people will beleive it. Manifestos don’t lie.

  16. @Ian S,

    Chris Grayling is rather cute…. :) :)

  17. Surdoc
    And independent and independant……..

    Typo is one thing, not spelling correctly is another. David Cameron, despite his denigrated education, will *never* make those mistakes.
    Roland
    I thought my point was clear. I think these pollster people, for whom I have great respect, should have given us their reasons. In a much earlier post I wondered whether it was the lazy so-called quality press that had left these out.

    My winnings on the horse were the smiles in the bosom of my family (well me and her) – I don’t gamble but will feel the same pleasure when my 36 34 22 charges home .

    BTW my ‘me and her’ was purposely colloquial. I do not make mistakes in English and we all should try harder. :-)

  18. Rob,

    As it happens, I was agreeing with you.

    Shy voter adjustment significantly reduces the Conservative lead, with ICM and Populus (take a look at their datasets).

    Surbiton, it is perfectly reasonable to look at past elections, to have an idea of what would happen at the present one. You said it was mathematically impossible for the Conservatives to win a 40/50 seat majority. Mathematically, you are wrong.

  19. In order for Labour to lose their majority, a grand total of 7,418 voters in Labour’s most marginal 24 seats have to change their vote from Labour to the closest challenger.

    26,406,050 people voted in Great Britain at the last election in total.

    7,418 out of 26,406,050 = 0.028%.

  20. @ Sean

    Erm, yes I realise what an exit poll is! My mention was because I was drawing attention to the fact that they became very accurate after methodology changes and yes, they have much larger samples to give a +/- 1% accuracy. Is that not what has been done with opinion polls too in order to improve accuracy, and therefore why do we (including polsters) not believe them to predict the right result? A poll of polls should statistically produce a result of similar accuracy given the sample size.

  21. I cannot understand why there is still this lack of understanding about what the group of pollsters meant when they unaminously gave their personal predictions about the GE result.

    It has been said several times before and is surely self-evident from the first posting of this thread. They were asked not what their current polls indicated would be the result, but rather what their personal predictions were for the actual GE result. In other words to crystal-ball gaze.

    On current polls I think most fair-minded people would say the polls at this moment, if replicated exactly at the GE would indicate a hung parliament with the Cons as the biggest party.

    Clearly these pollsters feel that events of the next 3 weeks are more likely to result in an increase in the Con lead leading to a small Con majority. There are several events such as the debates, immigration, economic figures and the manifestos etc. which are likely to have some affect on voting intentions. The pollsters think such future events are more likely to favour the Cons, but only by a fraction. That may nor may not happen. Their guess is as good as anybody’s at the moment.

    They are all hoping their predictions are accurate and that their polling method picks up any such trend in whatever direction, when/if it does occur.

    So some of you should stop trying to rip into them for their predictions in this very finely balanced situation and accept they too are allowed to speculate. If they are not trying to justify their predictions from currrent polls, why should you worry. Too many of you are too fond of making predictions yourselves about the future and presenting them as either fact or inevitable.

  22. It’s a bit of a myth that all of the exit polls in 1992 were completely wrong.

    ITN’s exit poll, for example, was: C – 41%, Lab – 37%, LD – 18%.

    So they were only out by 2% for C and Lab, within the margin of error, and they were spot-on with the LD share.

  23. Andy JS

    Not for the first time today, you have boggled my mind. my gast is flabbered.

  24. Sorry, you posted so quickly Andy – I meant the 7,418 voters who lose the Government’s majority.

    What number give the Tories a majority of one seat I wonder?

  25. EOIN CLARKE
    A Quattro is the Audi brand name for its four wheel drive system which most of the more expensive Audi’s have. The original was a very quick sports coupe from the 70s and 80s as per the Jam Jar (car) used by Gene Hunt aka Dave Cameron. That model is seen as an utter classic like Steve Mcqueen’s Mustang.

    Tune in after the GE for mor interesting car facts.

  26. Neil A – “To those who ask why, if the pollsters think their own polls are wrong, they don’t just alter them…” I don’t think anyone’s suggesting that, but if they think their methodology is off, wouldn’t they want to tweak it?

    FrankG – But surely on a site dedicated to analysing trends, it was always going to cause raised eyebrows that ALL of them predicted a Tory win. Sure they are entitled to their opinion, but it is rather odd.

  27. It is interesting to note that 4 of Labours 24 most vulnerable seats are to the LibDems not the Cons. It would only take the Labour share of the vote to drop by 3% and the LibDem vote to be the same as in the last GE for all these four seats to fall. If there is only a 1.5% swing from Labour to LibDem in these seats then these will fall. Depriving Labour of its majority is not only about Cons taking Labour seats, any party takiing Labour seats will do.

  28. Just to ask about the YG ‘correction’ recently which has led to YG being in-line with others rather than showing a narrower tory lead.

    It seemed to me that this ‘correction’ may have come almost immediately after the Telegraph criticism of YG, despite the protestations of PK.

  29. I don’t find anything extraordinary about 8 top pollsters believing that the Tories will slightly overperform their current polling numbers.
    However, I do find it a little strange that senior pollsters can be so confident at this early stage that the polls will still be showing a similar Tory lead in 3 1/2 weeks time. Particulaly, as the polls have listed between a Tory lead of 5% and 10%.
    What would greatly interest me, however, is what those pollsters believe is the minimum national percentage lead the Tories need to secure an overall majority. Is it 6%, 7% or 8%?

  30. @HOWARD
    I must have misunderstood you, I dont think however, these people will give any reasons until after the obese female has vocalised.

  31. Any polls out tonight – I know it’s Sunday but we had 3 last week if memory serves me correctly.

  32. @JulianGilbert

    “@EOIN – “If George Osborne can guarantee no VAT rise, he will be the next Chancellor.”
    People wouldn’t necessarily believe his guarantee though.”

    I suspect Eoin was being mischievous. He seems particularly frisky today.

  33. Sue – I’m curious as to why you think it “rather odd” that all the pollsters (currently) predict a Tory win. I can see how it would be displeasing to non- Tories but not how it would be “odd”, particularly as Sean – despite the incredibly disrespectful comments directed at him – has outlined what these predictions are likely to be based on.

    The alternatives are:

    – all eight pollsters are clueless; or

    – they are all deliberately making false predictions as part of a vast cosmic Conservative conspiracy to depress Labour activists

    Neither of these seems at all likely to me.

  34. “Suzy Quatro is a great 70s Rock Star. OnePoll’s male voters say she has the nicest celeb bottom, just a head of Chris Grayling.”

    Or Chris Grayling’s head with the body of Suzy Quatro – there’s a picture

  35. @JAMES LUDLOW
    As a caring kinda guy, rather like one of Labours recent leaders, I have been asking Labour supporters to control their hopes due to the distress of unfulfilled dreams. I think this is the first instalment.

  36. @ Sue Marsh
    I thought the site was “for non-partisan discussion of elections and polls”. The only poll than matters is the one on the 6th May. These pollsters were not giving an opinion on this site, but were crystal-ball gazing. I must say that their conclusions do not match my own interpretation or wishes. it is generally thought that DC will be better in the debates. It is generally thought that immigration will hurt Labour more. it is possible that future economic statistics due out before the GE may be better, but it is arguable if this will help Labour or merely reinforce that we can actually afford the Cons reversal of the Labour NI increase. Overall the ‘currently’ foreseeable future events seem to be more favourable to the Cons. Hence it is perfectly reasonable to expect their poll postion to either strengthen or increase. That is all these experts are telling us.

  37. Looking at YouGov’s regional analysis (and accepting that the regional samples are to small to read too much into) the Conservatives appear to be picking up most in the South and London, where the Lib Dems did wel last time and actually losing out in Scotland, where the Cons and Lib Dems have both done poorly in recent years and Labour extremely well.
    This might appear to be consistent with:
    No real swing whatsoever from Lab to Con but
    A small swing from Lab to Lib Dem
    A ferocious swing (relative to past Lib Dem numbers from Lib Dem to Con
    A small swing from Con to SNP, largely tactical voting.
    If (and it’s a big if) this is correct, it could very, very bad news indeed for Vince Cable in Twickenham, where he has only a few Lab voters to go out and squeeze. Remember, a 5.5% swing relative to all voters is a 25% swing relative to Lib Dem voters.
    It will, however, come as a relief to a lot of Northern English Labour MPs, though admittedly not to their new candidate in Liverpool, Wavertree.
    Are the Lib Dems in for a bad night, with the SNP and Plaid Cymru ending up as kingmakers?

  38. @James Ludlow
    I understood the doubts to be that these firms are mainly in the service of Tory supporting newspapers and other companies, the charge being that whilst they will stoutly defend the veracity of their polling, no such stricture would apply to their speculations. My point to Roland and the conspiracy theorists is that we have neither evidence for the charge of kowtow nor any of their evidence to justify their speculation.

  39. @Eoin

    If Cameron wants this election, one simple maneouvre would secure it. Labour have not included a pledge not to put up VAT.
    A gaping hole emerges right in the heart of their election strategy. If George Osborne can guarantee no VAT rise, he will be the next Chancellor.

    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

    With reference to VAT I have just picked this up of the BBC news web site

    “Labour’s election manifesto will promise not to increase income tax – but there will be no such pledge on raising VAT, the BBC has learned.”

    Well half your prediction has come true.

  40. James Ludlow – Surely you are just toying with me?
    The poll leads since the start of the campaign (and months before) have all been somewhere between 5% and 10%. The marginal effect doesn’t seem to be as strong as previously thought.
    Almost none of the polls have given much likelihood of a Tory majority.
    YET, not just a majority, but ALL 8 pollsters predicted a Tory majority.
    I never post blindly partisan claptrap, I rarely choose to believe polls I “like” over what is probably the real story, so I believe the points I’ve listed above back up my feeling that it is indeed “odd”. To give you an example, I would also think it “odd” if tomorrow 8 polls came out all putting the Cons, Lab and LibDems on the same figures. It is statistically very unlikely.
    But don’t trouble yourself, I don’t ask you to agree!

  41. @Greengrass
    I’m glad to say I’m not that good with photoshop.

  42. Ian S
    Were those remarks applicable to the entire country (no swing Lab to Con) because if so, I could not follow you?

  43. Rosie p – coincidence. It was going to happen anyway, as it did in 2005. Peter and I decided to make to switch at the beginning of the campaign rather than just the last poll a couple of months back, and had been testing it for a couple of weeks before the silly Telegraph hatchet job came out.

  44. Incidentally, perhaps those constantly posting to me on this, think I see some sort of “conspiracy”??
    Not a bit of it. Sean Fear gave me a perfectly reasonable suggestion that I accepted.

  45. @ Howard – good grief! So how do the conspiracy theorists explain ICM and ComRes going along with this diabolical masterplan?

  46. @ Sue – no, I’m not toying with you at all. I’m genuinely curious. My assumption in such a scenario would be simply that the pollsters know a lot more than I do about polls and political trends, them being pollsters and me being a photojournalist.

  47. @Howard
    A good question
    Yes, I mean in the country as a whole. I think the figures are actually consistent with not one single Labour voter from 2005 going out and voting Conservative this time. The Conservatives could be getting all their extra votes from the Lib Dems.
    I’m not saying it’s so. I just think it could be consistent with the figures.
    Willing to be refuted if it’s not consistent.
    Must go to dinner, back later.

  48. James L
    Well, I’ll speculate for them. I gather that the two you mentioned worked for what you consider are lefty papers. Well, the vast majority of all pollsters’earnings come from marketing companies and the like (source AW) so the rest is obvious (if you are a conspiracy theorist).

  49. On a light note has anyone seen the new UKIP poster. It shows a photo of GB DC and NC witha slogan underneath saying
    “sod the lot”. Nice.

  50. Roland – Is that one of your jokes? If not, it might just be the most successful poster of the campaign :)

1 5 6 7 8 9 11