It’s very early, but we already have the figures for tomorrow’s YouGov/People poll, courtesy of their political editor Nigel Nelson’s twitter feed. The topline figures, with changes from YouGov’s last poll, are CON 40%(nc), LAB 28%(-3), LDEM 18%(+2).

While it shows a drop for Labour, this poll is very much a reverse from YouGov’s last – the one straight after the Pre-Budget Report which rather oddly showed a negative reaction to the PBR, but Labour nevertheless gaining four points. This largely reverses that movement.

In tomorrow’s papers I’m also expecting MORI’s “latest” monthly poll, though it would actually have been conducted a week ago.


60 Responses to “YouGov has Tory lead back to 12 points”

1 2
  1. tory vote seems very stable of late, while the labour vote seems all over the place. Is this due to Labour undecideds shifting between the lib dems and others or is there some other reason?

  2. IMORI has rather lost its glitter at this late stage. If, as some have suggested its in Tory single figure lead country, then this You Gov will be more up to date and therefore more meaningful.
    Those of us who have talked about Tory 11 to 15% leads look about right, especially if the LD’s are getting people back from Labour. As per usual Mr Cable spoke quite well after the PBR
    and may well have turned a few heads.

  3. So still happily sitting in the 40/30/18 +/-3% that the other polls, EAR (Exluding Angus Reid) are suggesting.

    Can anyone who really thinks Angus Reid is a true reflection of voting intention, please explain to me why the pollsters that very accuratetely predicted the last election and Euro elections are now suddenly overstating Labour by 4-6%?

  4. As I have stated previously the consistent Conservative figure of 40% is astonishing and should not be happening if looked at statistically .
    If the true Conservative position were exactly 40% then a run of opinion polls should show because of sampling/Mof E and different methodologies a bell shaped plot of the measurements of the Conservative support . with a range from 37 to 43 % with perhaps an odd outler outside this range .
    Instead we are seeing almost a flat line . I have no suggestions as to why this should be simply pointing out that this should not be so .

  5. There is apparently a rumour that the Mori from last weekend has the lead down to 3.

    It was taken at the same time as the ICM which had to scale down the Labour responders to a huge extent.
    IF the rumours are true, no doubt there will be a big splash on that one poll tomorrow however much it flies in the face of the others with both Comres and YouGov showing bigger Tory leads and move back.

    I agree with the comment above. Mori needs to stop generating headlines by being at the extreme of polling and out of date.

  6. Stratto, I think you will find that the pollsters overstated Labour at the Euros.
    If memory serves it was around 2 to 6 %.

  7. Well, one thing at least seems certain. The Tories will have the largest share of the vote at 40% or so. That figure seems almost set in stone. Whether or not that 40% gets them a majority of seats is down entirely to the relative performance of their opponents, and the effects of tactical voting and “others” returning to the main parties.

  8. If the rumour is true and mori has a lead of 3 points, i’d expect it to be plastered over the observer tomorrow and they’ll ignore the fact its out of date.

  9. ‘T WHITE
    tory vote seems very stable of late, while the labour vote seems all over the place. Is this due to Labour undecideds shifting between the lib dems and others or is there some other reason?’

    Possibly the benefit of being opposition. Labour is doing things -rightly or wrongly- which impact on people. ‘if ‘bad’ people will in the main first thing of going to Lib Dems as it’s not as big a leap s going tory. As such, Tories 40% represent those there for a while. The other changes are more due to daily news swing…

  10. Sally C

    Please check your facts (button marked European Elections on the left). You Gov were 0.3% out, ICM 1.3%, Populus 0.3%.

    Only Comres (with its unusually unweighted final poll was way out, and as it was outside margin of error of the others, I would have rejected it as a rogue).

  11. WMA 40:28:18 so YouGov seem to be spot-on. Less than 140 days to a May election and the downtrend in the CLead over the last 140 days remains, although over the last month the CLead has remained flat.

    Seems unlikelt that the Copanhagen fiasco will have helped Brown, but we shall see…

    I wish I understood the false recall business. How do pollsters know how many people voted Labour but said they didn’t? It seems like a very odd fudge-factor to me. Can Anthony explain?

  12. @NBeale – “I wish I understood the false recall business. How do pollsters know how many people voted Labour but said they didn’t? It seems like a very odd fudge-factor to me. Can Anthony explain?”

    I’m also intrigued by this. My understanding is that fewer people admit to voting Labour in 2005 than actually did, so the 05 Labour voters are weighted up to account for this. What I’m really interested in is how accurately this reflects ‘shy Labour’ voters, if indeed there are any. AW might confirm this or otherwise, but I understood the false recall stuff was brought in to protect against another 92 style event where Tory support was underestimated. However, it seems this problem continued in the three successive elections and I’m interested as to whether there is now a false Labour issue as they are generally regarded as the unpopular party. This could explain Mark Seniors point about why the Tory support is level while labour’s bounces up and down – maybe this is the effect of the various approaches to dealing with this problem, and the problem itself?

  13. Alec,

    My understanding is that it is the other way around – more people say they voted Labour than actually did. So those who say they voted Labour are weighted down.

    Thus AR, who weight to actual 2005 result, weight down Labour by more (essentially ignoring false recall), while the others, who weight part way to the 2005 result, don’t weight Lab down by quite so much. MORI don’t weight by this at all, but their high likelihood to vote filter tends to counter this because Lab voters are less likely to vote.

    The Weighting link on the side bar explains all.

  14. It is important not to over emphasise the effect of adjustments to poll figures for false recall . It will not change the headline figures by more than 1% or 2% at the outside ( the latter only if the pollsters should be reversing the effect . Much larger adjustments are made because of past vote weighting particularly where the sample has a large pro Labour bias as for example the ICM poll had this month .
    We should also remember that M Of E can have a greater effect than actual movements in voter intentions . If the Conservatives are at 40% , the statistical chance of an opinion poll finding them at 40% is only around 30% , There is a greater chance of them given them an incorrect figure than a correct one although only in roughly 1 poll on 20 will it be out by more than 3% .

  15. All good points Mark. We hear a lot about the variablility of the polls, but the fact is that most (excluding AR) are within margin of error of the others.

    AR are consistently outside, or on the extreme lower fringes of the M o E for Lab, which is why I think it is not measuring the same thing as the others.

  16. Alec – as Statto says, those pollsters who weight by it assume it is the other way round. More people claim they voted Labour than actually did.

    ICM, Populus, etc base their decision to weight by past vote on the belief that it does not change vastly over time. This implies that most false recall is NOT a reflection of current support (i.e. people who voted for a party they no longer support saying they voted how they wish they had), or it would change as the polls change (MORI, I should add, think it could be a result of this, making it inappropriate for weighting).

    My guess is that the main factors are people not admitting that they didn’t actually vote, and instead claiming they voted Labour (this, at least, can be proved to be true in at least one case. The British Election Study actually validiated some of their data against marked electoral registers, so we know for certain that in the BES survey at least, people who claim they voted when they did not actually do so disproportionately claimed that they voted Labour). Secondly, I suspect people who voted tactically sometimes claim they voted for the party that would have been their first choice, not the one they actually voted for (in this case, increased the recalled level of Labour support and decreasing that for the Lib Dems).

    Phone polls continue to find too many Labour supporters and require weighting down, suggesting that is nothing to do with unpopularity and do to with attitudinal (i.e. willingness to answer a survey) or lifestyle (i.e. availability to pick up the phone for a pollster) factors. The adjustment for don’t knows however (the so-called “shy Tory adjustment”), which in the 1992-1997 Parliament favoured the Conservative party has long since reversed and has helped Labour for many years now (Andrew Cooper of Populus rechristened it “Bashful Blairites”, though I suppose they must be Bashful Brownites now).

  17. @ Mark Senor

    One possible theory is that those who have decided to vote Tory have pretty much made their minds up, which is why the Tories are stable on about 40%.

  18. @T White

    “tory vote seems very stable of late, while the labour vote seems all over the place. Is this due to Labour undecideds shifting between the lib dems and others or is there some other reason?”

    I totally agree with your comments. I have been saying this for some time but have been shot down by several people on this site.

  19. Anthony (& Statto) – thanks for clearing up my confusion. Now you’ve explained eveything we can get on with the business of being mystified why one poll says 17%, one says 12%, one says 9% etc etc….

  20. Mark R , I have not argued either for or against the Conservatives being at a steady 40% in the polls , I have simply commented that statistically if they were at 40% , then 7 out of the last 10 polls should not have shown them at 40% . They should have ranged from 37 to 43% with a roughly 50% chance of one outlier .

  21. @ Statto 5.10pm

    ‘AR are consistently outside, or on the extreme lower fringes of the M o E for Lab, which is why I think it is not measuring the same thing as the others.’

    That seems a wholly reasonable viewpoint. However, is it determined precisely what the various polls are seeking to measure? For if not just AR, who are commissioned by Political Betiting, but all the others have diverse aims, we need a clear statement of what and to whom each poll is answering. And if it changes its interlocutor from one poll to the next.

    Perhaps every poll puts out a statement about what it is doing and for whom before it publishes each set of results?

  22. Mark Senior

    Presumably 40 means >= 39.5 < 40.5. I guess 7/10 in this range is unusual, though perhaps not outrageously so – anyone fancy doing the probability calculation?

    Certainly the Conservative share shows much less variance than either Lab or LD. I can't think of any statistical reason why this should be so.

  23. Agree with Mark Senor that 40% of voters have decided on the Tories. I cannot see this being reversed as it has been reasonably consistant at this level.

    Once a solid base of 40% becomes even more consistant variations I believe will be small and insignificant.

    A hard core Tory support is emerging and this will be very difficult to disturb.

  24. Mark R,

    Perhaps some of the 40% are scored when the real score should have been a little higher or lower?
    Still the phenonmenon allows everyone on here regardless of tendency to agree the cons are indeed around 40%.
    I would suggest in fact a little higher as some others will return.
    For me 40% will be enough for a modest majority, even 38% will probably do which gives the cons a decent cushion.

  25. @Hatfield Girl

    Certainly all the polls have different questions, but presumably they all seek to measure voting intention?

    AR’s “support” question is the most marked difference. I really can’t understand why they chose this term (does it have different connotations in Canada?).

    I believe their stated aim was to try to measure tactical voting, which would have been a really useful and interesting perspective. The “in your constituency” part of the question might help do this. But that goal, in my view, has not been undermined through the use of the term “support”.

  26. Mark Senior’s comment makes no sense at all to me. There’s absolutely no reason why poll after poll shouldn’t show the Tories on 40% if that reflects the way people are responding to polling questions.

    With this YouGov poll it’s interesting that the Labour vote goes down and the Others vote goes up. For some reason a lot of people don’t seem to like facing up to the fact that quite a lot of Labour voters are going to register a protest against the government by voting for independents, UKIP, BNP, Green, etc. It seems like quite a likely thing to happen in my view given that the government has failed in some of the key policies which Labour governments are supposed to support, such as reducing the gap between rich and poor for example.

  27. Sorry wrong Mark , shoud have been to Mr Senior.

  28. Andy , if I were a waelthy person I would demonstrate my point quite easily .
    I would commission ICM to conduct 20 identical polls at the same time . If the Conservative support was at 40% , you would statistically expect roughly 7 polls to have them at 40% , 3 each at 39 and 41% 2 each at 38 and 42% , 1 each at 37 and 43 % with 1 outlier outside that range .
    If I was being mischievous , I would have them published in order showing the Conservative support steadily falling from 43 to 37% and watch all sorts of theories being put forward as to why this should be happening when of course the reality would be that all the polls were variations measuring the same level of support of 40% .

  29. If the BBC makes a song and dance about a possible 3% lead in this week-old MORI survey but ignores the latest YouGov poll the proposition that they’re utterly biased against the Tories will have been proved once and for all.

  30. @Mark

    There is no “40 Set In Stone” figure in the polls.

    Looking over the last 20 polls, we see a range of polls from 44 to 37, with the high points slowly moving down. The last 10 polls, the high point as been 41. But 40 has repeated quite a lot.

    And people like to find patterns, and like round numbers, and going below 40 is a psychological barrier for a certain group of people here.

    That said, my moving average is currently at 40.16, having moved back up over 40 after sitting just below it for some time. Could be a levelling off, or could just be the effect that the conservative-leaning polls came out later than the labour-leaning polls and have just given it a bit of upward ‘correction’ on an otherwise downward trend.

  31. Will there be another poll published tonight ? Or, is this MORI poll a big hoax ? I just wonder if MORI would do any fieldwork without a buyer !

  32. Surbiton – My understanding is that Ipsos MORI always do their monthly monitor without a client.

    They used to have a contract with the Times, up until 2001 or so. After that contract ended they continued doing the monthly monitor themselves so they was no gap in the data, sometimes finding a client to publish it, sometimes just press releasing it themselves.

  33. ANDY STIDWELL makes a very valid point. If the BBC do use the 3% figure they will confirm their support for the Labour party. They would do better to look at the statistics on this web page for realistic figures and comments.

    If the 3% figure\is correct it will raise serious doubts about the credibility of the Mori survey as it will be way outside the recent trend.

  34. Ipsos Mori always do the fieldwork for their monthly monitor without a buyer . They then usually try to sell the poll results with the Observer being the most frquent buyer but some months they do not generate any interest in buying the results and will publish them themselves .

  35. @Andy Stidwill – “If the BBC makes a song and dance about a possible 3% lead in this week-old MORI survey but ignores the latest YouGov poll the proposition that they’re utterly biased against the Tories will have been proved once and for all.”

    I’m afraid it would do no such thing. Two points;
    1) The media isn’t brilliant at polls, statistics or science. If you want detailed analysis of such things go somewhere else. What they are good at is news. ‘Poll shows same lead as everyone else’ isn’t news. ‘Poll shows dramatic drop in lead’ is. never mind that they might gloss over the intricate analyses of all recent polls etc – right or wrong, a 3% is news.
    2) You really have no understanding whatsoever of the legal strictures that are in place on the BBC (and all TV broadcasters in the UK). There is no, and can be no, bias against one or other party. I find it frankly laughable that anyone thinks the BBC is biased against the Tories (or anyone else for that matter). The BBC output is monitored and analysed in minute detail and I’m afraid there is no general bias in their output whatsoever. If you really want to understand what TV bias looks like, try watching Italian TV or Fox News. Then come back and try to tell me the BBC is biased.

  36. I’ve always suspected that the BBC lean towards the Labour Party probably because Labour have filled it with their cronies like every other public body which is why we have this continuous PC C##P pushed in our faces at every turn!
    David Cameron must remove them at the earliest possibility as they are causing untold damage to our social stucture!
    Almost everyone you speak to is sick to the back teeth of whats going on in this country.

    The latest poll simply confirms that most people seem to have made up their minds although I feel sure there is a very dirty Election campaign looming straight after Christmas/New Year!

  37. Statto 6.22pm.

    Thank you for that; your remark, ‘presumably they all seek to measure voting intention?’ is what is being called into question by discussion of all sorts of things other than arcane statistical phenomena. There seems to be a view that some polls are being used for other purposes. Are arcane statistical phenomena being manipulated to produce superficially neutral data on shifts in political sentiment, and thus to produce those shifts?

    A ‘poll statement’ prefacing each poll, giving information on the poll commissioner, the information they sought, and statistical techniques employed, would move polls back into being data suppliers, and out of being just an esoteric part of general political analysis.

    Telling it like it is, is much more interesting than telling it like the poll commissioner wants to hear, or wants to use under a cloak of polling ‘neutrality’.

  38. Don’t be so ridiculous. Nobody in the Labour Party sees the BBC as biased in their favour. And if Mark could actually name the “cronies” he blathers on about I will be extremely surprised & even maybe impressed.

    One thing is for sure. If Tories have nothing better to do than complain that the BBC hates them & is full of closet Pinkos & Marxists, the British people will become pretty biased against the Tories.

  39. Hatfield girl – newspapers are all supposed to print a blurb along the lines of “Polling Company Ltd interviewed x number of people between the dates of y and z. Polling Company Ltd are members of the BPC and abide by their rules. More information is available at http://www.pollingcompanyltd.com” and then have detailed tables and of their methodology somewhere on the website.

  40. Well Barnaby we had better just agree to disagree because thats my view and I’m not wasting any more space boring everyone on this thread!!

  41. I would not say the bbc are specifically bias. It is true that they have more left liberals than right wing people in there organisation

  42. Anthony

    Any news of this Mori poll yet?

    When do you expect it to break and have you any inside knowledge of the figures?

  43. Big John – no idea, whenever the Observer put it on their website (assuming, that is, that it is in the Observer). And no, I don’t know anything at all about it.

  44. Lots of tories think the BBC has a labour bias, lots of labs think the BBC has a cons bias. Both claims are absolute bollocks.

  45. Rumour says, the “Observer” poll will be released at 9:10 pm. I make it 9:14 now.

  46. according to Sky News the Tories have a 17% lead in the Observer poll. That is all that they said, no share figures given.

  47. The Sky News website says no such thing.

  48. The PA is now reporting 17% Tory lead in the Observer. It is a MORI poll.

  49. Frenzied rumours on PoliticalBetting regarding the MORI poll. Most of the day people have been predicting a 3% lead and in the last few minutes it’s changed to a 17% lead. Don’t know what to believe!

  50. Ipsos Mori Observer
    Con 43 Lab 26 LibDem 20

    All rumours were incorrect

    Conservative supporters will vhange from saying how rubbish Mori are to how wonderful pollsters they are in the blinking of an eye .

1 2