Tonight’s YouGov poll has topline figures of CON 40%, LAB 38%, LDEM 10% – full tabs here. This certainly suggests that the five point Labour lead yesterday was indeed an outlier, and that the underlying picture remains the same old neck-and-neck position between Labour and Conservatives.

372 Responses to “YouGov/Sun – CON 40%, LAB 38%, LDEM 10%”

1 2 3 8
  1. What a day of turnups. Harry gets off. Fabio walks out. The Tories’ goose is not cooked!

  2. I see the pro Ken advertising in this site isn’t getting any less crude.

  3. @Neil A

    “What a day of turnups”

    When I first read this quickly I thought you said turnips and I wondered what Graham Taylor had got to do with the conversation!

  4. @ Neil A

    Did you see this in the DM ? The Police Officer chasing himself.

  5. What a complete joke YG has become and so utterly predictable. I could have bet my house on a result like this. Sorry but my trust in YG has completely evaporated, no credibility at all,

  6. YouGov put the LIb Dems in double figures again. At the risk of repeating myself this is much more in line with what the other pollsters are telling us (apart from YouGov eight pollsters now put the Lib Dems in the range 10-16%). I also think it’s very interesting that, according to YouGov, 22% of voters put a coalition involving the Lib Dems as their first choice for the government of the country. This is the pool of pluralist, pro-coalition voters who the Lib Dems need to persuade back into the fold.

  7. @R Huckle,

    Yup I read it this morning. It didn’t surprise me – knuckleheaded doesn’t begin to describe some of my colleagues….

  8. @Gracie,

    It looks like major variation, but it is all well within MOE around a small Labour lead. Really nothing to get in a tizz about.

  9. @R Huckle

    “Did you see this in the DM ? The Police Officer chasing himself.”

    Maybe this is Theresa May’s big idea to increase productivity and reduce costs. They could be stopping and searching themselves soon, although maybe that will be confined to police officers from ethnic minority backgrounds. lol

  10. @Neil.

    Nope I am not in a tizz, in fact I am “untizzed” as “untizzed” could be :0)

    I just expected this, it has been the pattern for weeks now, ever since YG started to use their new methodology.

  11. A Detective colleague of mine retrained as a barrister and left the police. For his leaving do I prepared a set of mock case papers in which he arrested himself for an offence and then represented himself in the interview he conducted with himself.

    The interview transcript was one of my finest, surreal creations.

  12. NEIL A

    That sounds really funny! :-)

  13. 2010 vote looks a bit iffy again.

    Low on both Tory and Labour side and high in Lib Dem and Other side of the equation.

    Con 35% (-1) Lab 26 (-3) Lib 25% (+2) Other 15% (+3)

  14. @Richard O

    [Crossbat (and indeed, everyone else) – don’t rise to partisan comments, just ignore them and I’ll delete them when I see them… AW]

  15. Well whatever he was today he certainly caused cameron to have an attack of the crimson willies (metaphorically speaking of course).

  16. My money’s on Mike Bassett. Why would Rednapp want the job?

  17. Gracie –

    Erm, the YouGov methodology change took effect on the 17th October, when Labour had a lead of around about 4 points. Labour’s lead then grew a bit to 5 or 6 points throughout November, and then narrowed to roughtly neck-and-neck after the European “veto” in early December. In other words, if the poll narrowing was due to the methodology change it mysteriously took 7 weeks to have any effect.

    More to the point, the narrowing in the polls has been echoed by every single polling company, it is hardly a strange YouGov thing. YouGov have, on average, been showing a Labour lead of 1 point in the last week or so. Looking at the last polls of other companies, they are

    ICM 5 pt Con lead
    Opinium 2 pt Con lead
    Ipsos MORI Con and Lab tied
    Populus 1 pt Lab lead
    ComRes 1 pt Lab lead
    Angus Reid 2 pt Lab lead
    TNS-BMRB 3 pt Lab lead

  18. Anthony,

    It’s very generous of you to dignify “Gracie’s” partisan outbursts with a reasoned response – well done.

  19. Just back from seeing Sunderland rob the Boro.
    That’s my partisan bit out of the way.
    Neck and Neck, small lab lead maybe but not with ATTUK. Cons will be most pleased with polls recently but Lab not disheartened either.

  20. Having BOTH major parties up from last GE is clearly a UK exception. If one takes a look at the polls for EU countries that face GE In 2012-13, one sees two tendencies:
    1. Both major parties are down, especially when they govern together or support together a technical government (Italy). This is the case in Greece, Italy, Lithuania and Austria.
    2. The senior gvt party is down and the major opp. party is up. This is the case in France, Slovakia, Romania, Malta and Bulgaria – in the firstl four cases social democratic opposition is clearly ahead of center right gvt, sometimes dramatically so, as in Romania and Slovakia (in latest Slovak GE VI poll, the party of incumbent PM has fallen from 15 to 5% and from 2nd 6th place, barely clearing the threshold for entering Parliament).
    The only country where one can observe a pattern similar to that of UK is Germany. According to latest polls, Merkel’s CDU is at 35-38 (GE 2009: 34) and opposition SPD at 28-30 (2009: 23). The difference with UK, though, is that Merkel’s liberal allies (FDP) have lost all hope to enter next Parliament (average 3%, from 14.5 in 2009), so the incumbent coalition is doomed anyway.

  21. 30-poll comparison (most recent 30 polls compared with previous 30 polls : simple averaging):


    Con – Up everywhere expect London;
    Lab – Down everywhere expect London;
    Lib – Down everywhere expect MIdlands & Wales;
    UKIP – Down everywhere;
    Green – MIxed bag;
    SNP – Up.

  22. The two party share will probably be 74-75% next time (GB ),
    with the LDs on 15-16%, and the others 10-11%.

  23. JJB – I agree.

  24. @all

    Was I the only one who saw that police officer story and thought hold on a minute, I’ve seen that plot before…


    As ever, thank you.

    Regards, Martyn

  25. @Anthony

    What Sergio said


    Thank you

    Regards, Martyn

  26. @Virgilio,

    Interesting update, as ever.

    Your figures for CDU vs SPD look like Merkel is doing a bit better than she had been, or am I misremembering? I thought that the CDU had been getting hammered in 2011 in state polls?

  27. I’m less sure how many seats the LDs will lose.
    It’s still likely to be somewhat hit and miss – although could be fairly subsantial.
    They are still able to buck trends and inflict upsets, although the May 2011 results do rather suggest that unless one of the other 2 parties is doing badly (which didn’t happen then) they will lose substantial numbers of seats if they go below about 18%.

  28. @Statgeek

    “Green – Mixed bag”

    I don’t know about that. I for one am pretty excited to be consistently polling one or two points over what we got in the last election but I guess minnow parties are easily impressed :D

  29. @AW

    A seven-point swing in a day. Whilst I know this is within the MoE, it nevertheless emphasises that any single poll, in isolation, isn’t worth the paper the calculations are written on.

    Given that the parties clearly are in neck-and-neck country, might it be worth increasing the sample size significantly; say double it. I realise that it will cost more, and accuracy won’t increase proportionately (if my memory of A-level statistics 4 decades ago is correct, it’s a root function), but the benefits of stability and reliability may overcome this.

    But you know both the facts and the finances, so I may well be whistling in the dark here.

  30. @ Sergio what partisan outburst? I have a beef with YG methodology and said so and Anthony has explained it, although I am not entirely convinced I thank him for doing so.

    Anthony, but isn’t it true that almost all of the other major polls methodology either favour the Tories or the Lib Dems? (and this is not a partisan remark) the truth is that soon after YG changed their methodology the Labour lead started to dissipate? There may have been a few results at 4-5 points etc, but these themselves could have been outliers.

    By the way this is definitely not partisan, I would have mentioned it if it had been the other way around as I don’t much go in for fooling myself.


    If anyone was actually interested in polls which accurately identified likely polling, the money would be better spent in running demographically balanced polls in the nations/regions in GB – they might even consider the detached NI too, but that is a vain hope.

    If YG has a large enough panel in Scotland to run three simultaneous demographically balanced polls in Scotland of around 1000 each, then they could (with little extra effort) provide realistic polling across the UK or GB.

    Their clients, of course, don’t want such detail – that would detract from their narrative that there is any such thing as “British” political opinion which can be measured in the crude way that they prefer.

  32. @ Oldnat – Your post makes sense, more sense than suddenly “guessing” that voters who say they don’t know will revert to who they voted for last time anyway.

  33. Gracie,

    “the truth is that soon after YG changed their methodology the Labour lead started to dissipate? ”

    Anthony has already addressed your hypothesis if you read his post…….

  34. I may be a bit naive, but isn’t it true that these ‘rogue polls’ or ‘outliers’ are simply caused by having too small a sample. Why not have fewer polls with a greater sample. Then we would have a high probability that the result was accurate and could stop this nonsense.

  35. Gracie

    There’s various different things going on here. When YouGov changed their methodology it was to correct the newspaper weighting (nobody likes these but no one can think of a better alternative at the moment). It’s generally thought that previously these gave Labour a 1 point or so advantage by over-estimating the Mirror/Record readership.

    Meanwhile many of the other polls effectively apply some sort of ‘shy’ correction. At the moment this helps especially the Lib Dems because their percentage vote has dropped so much since May 2010. Maybe a small gain for the Conservatives, but not significant.

    In addition most other polls use some sort of turnout adjustment though YouGov doesn’t. This tends to help the Conservatives in the other polls, though which pollster is “right” is more a matter for philosophy than psephology.

    So basically clear as mud. Mainly because in addition a lot of voters aren’t sure at the moment which (if any) way to go.


    Time to look at Anthony’s list of common errors about polling, I think.

    Increasing the sample size of a coherent population produces a minimal effect on accuracy.

    At the same time, concatenating very different populations into a single sample reduces accuracy.

  37. Steve Coberman

    YouGov is currently running with samples around the 1,700 mark, but when you’re looking at the ‘headline’ voting intention (VI) the non-voters mean this is reduced to around 1300. Even in the best circumstances this would give you a margin of error of +/- 2.66, because pollsters have to make adjustments to get their sample near to a ‘perfect’ one, you’re best rounding that up to 3.

    Even then about one in twenty samples will fall outside that limit[1], but a change in Labour of -5 points or of Conservatives of +3 is well within the +/-3 limits. Remember that the Margin of Error applies to the actual value of VI – the difference will vary around twice as much (because when one goes up…)

    Doubling the sample wouldn’t help that much – as you say it’s not a linear function. YouGov ran the polls for a while with a sample of around 3,000 but that only gives you a ‘perfect’ MoE of +/- 1.74 say 2. It’s a lot of extra work and cost for not a lot more accuracy – where it does help is if you want to split your sample or look at sub-samples[2]. The difference between 500 and 1,000 is much more than between 1,000 and 2,000.

    [1] It’s actually far more complicated than that but it’s too late and I’m feeling too stupid to go into details.

    [2] Not that this stops various commenters on here predicting the outcome of the next election on a sub-sample that that turns out to be a drunk teenager in Peckham or an elderly lady in Arbroath with a malfunctioning mouse. But whatever.

  38. @ Colin (from the last thread)

    “Do you think “indulging in homosexuality” , rather than -say”practising homosexuality” , carries any connotation which might be seen as offensive?

    Do you think that “riddled with homosexuality” is a reasonable way of saying “many people are homosexual in that particular group” ?

    I only ask-not being homosexual , I find it difficult to judge.”

    I find the word “homosexual” to be offensive in the same way that many black people find “negro” to be offensive. Read Judge Reinhardt’s Prop 8 decision yesterday, he used “gay” and “lesbian”, not “homosexual.” Even the conservative Republican Mormon judge who dissented used “gay” and “lesbian”, not “homosexual.” Even John Roberts doesn’t use the term “homosexual.”

    Of course, you guys feel differently. It’s probably a cultural thing.

    I think “riddled with homosexuality,” “indulging in homosexuality”, and “practicing homosexuality” are all fairly offensive but it really depends on the context in which they’re said. Given the context of the statement, I don’t think it was homophobic.

    @ Neil A (from the last thread)

    “For once, I’ll defend Ken. I don’t think he’s remotely homophobic. And I don’t think his remarks about gay MPs were particularly offensive. At worst he can be accused of putting an unnecessarily partisan slant onto the issue.”

    I don’t think he was being homophobic either. However, I doubt that all gay Labour MPs are out. There are more openly gay Tory MPs than openly gay Labour MPs.

  39. OldNat

    It wasn’t YouGov who did the 3 X 1000 polling for Lord Ashcroft. It is ‘believed’ to to be Populus. Of course by doing it this way there is no obligation on Lord A to show all the details or within a certain time of initial publication. A bit like off shoring your polls to Belize.


    It may be “believed” to be Populus, but the tables published by Ashcroft are precisely in YouGov format, and when I asked Anthony about the lack of “Don’t Know” answers, he was able to tell me that the questions had been asked in that closed question format.

    Ashcroft uses both Populus and YouGov (perhaps others). it could even be that he pays VAT on some commissioned polls!

  41. SoCalLiberal

    ‘Homosexual’ is still seen as a neutral term in the UK – perhaps more formal than ‘gay’, but also a touch more inclusive as ‘gay’ tends to imply male. I suspect that the antipathy to the term in America stems from the ‘medicalisation’ of homosexuality in the 50s and 60s[1]. This never happened to the same extent in the UK so the term was never contaminated.

    Mind you if you’re going to talk about whether you pronounce the first ‘o’ long or short, there’s probably blood on the floor. :D

    [1] But then, let’s face it, what doesn’t get ‘medicalised’ in the US.


    ” ‘gay’ tends to imply male” – my niece’s daughter would find that comment discriminatory.

  43. @ Old Nat

    “” ‘gay’ tends to imply male” – my niece’s daughter would find that comment discriminatory.”

    There seems to be a small division between east and west on this where westerners are more likely to use the term “gay woman” and easterners are more likely to use the term “lesbian.”

  44. Have people seen the Martin Kettle article, in the Guardian:

  45. The LDs have been doing considerably better in local council by-elections recently compared to the opinion polls, although I know that a lot of commentators don’t think local by-elections are as a good a measure of support as they used to in previous times.

  46. @ Roger Mexico

    “Mind you if you’re going to talk about whether you pronounce the first ‘o’ long or short, there’s probably blood on the floor.”

    Is this something that would cause Elton John and Rupert Everett to start brawling with each other?

    As for the word, homosexual, this clip might prove your point.

    Or this one:

    h ttp://

    Here’s the long version of the first clip if you feel like watching:

    h ttp://

    And just a warning, this video is extreme offensive. Btw, Mike Wallace has never apologized for this peice or ever repudiated it. Frank Kameny, who’s featured in this video, has a street named after him in the District of Columbia.

  47. @ Roger Mexico

    Not that this stops various commenters on here predicting the outcome of the next election on a sub-sample that that turns out to be a drunk teenager in Peckham or an elderly lady in Arbroath with a malfunctioning mouse. But whatever.
    :razz: Nobody loves a [email protected] :-)

  48. *Feels unloved*

  49. @OldNat,

    Surely the fact that the LGBT community calls itself the LGBT community rather indicates that “gay” as a word is generally thought of as meaning a male homosexual.

    Otherwise “Lesbian, Gay, Bisexual, Transgender” would be a little bit tautological.

    Personally I think that too much effort goes into trying to define the “right” word or the “right” usage of a word for things. So long as people don’t use truly offensive language, there’s no reason why homosexual, gay, queer and any number of other words shouldn’t be acceptable.

  50. What effect will Capello’s resignation have on voting intention?

1 2 3 8