Tonight’s YouGov poll for the Sun has topline figures of CON 36%, LAB 41%, LDEM 11%. A five point Labour lead is larger than we’ve seen of late from YouGov (the last time they showed a lead that large was early February which, in hindsight, looked like a blip). Normal caveats apply: sure, it could be a sign of Labour’s lead growing beyond the point or two that YouGov have been averaging at of late, alternatively it could just be normal sample error.

There was also a Harris poll in the Metro this morning with some extremely odd figures. In the newspaper they are quoted as CON 32%, LAB 23%, LDEM 12%. There is no indication if the balance is others, don’t knows, won’t votes or whatever. I think I’ll come back to this one when the tables appear on Harris’s website.

UPDATE: The tabs for the Harris poll are here. The rather bonkers voting intention figures were, in fact, not really voting intention figures at all, but answers to a question on what party people were “most inclined to support” (the low percentages were, as many people guessed, because it included 19% who said “No party”).


121 Responses to “YouGov/Sun – CON 36%, LAB 41%, LDEM 11%”

1 2 3
  1. I don’t place any credence on the Harris (Interactive?) poll. The YouGov one is interesting but very different from the previous 2 days; we will clearly have to wait & see if it is corroborated before any real conclusions can be drawn from it.

  2. As for the polls tonight, I’m going for the Harris as the more likely of the two. A 9% Tory lead feels right to me and I’ve had a hunch for some time that Labour are in the 20-25% range, not the ridiculous 41%’s that we sometimes see in You Gov. When you take this with the ICM in January, I think we’re starting to see the Tory lead widening to something in the 5-9% range.

    Is Harris the new Gold Standard of polls, I wonder?

  3. 37% for Con+LD is around the average of recent You Govs so probably sample variation.
    Same edges of moe we have had for 2 weeks or so con 1% lead, lab 5% so no change really unless we see 2 more like it, suggesting 1-2% Lab lead.
    Don’t see any obvious reason why the lab lead would widen.

  4. I’ve been waiting for one like this from Yougov for about a week or so.

    They usually have at least one sample error a week and the 2010 vote looks far too low for the Tories as does the over 60 vote.

    I expect this to be a sample error to average out the mean as someone we all know might say after yesterdays Sunday Times poll.

    Could be wrong of course but we shall see for sure over the coming days before we can draw any conclusions.

    As for the Harris poll I think we can disregard that one as it’s clearly a load of nonesense.

  5. Interesting change in the tables. From the 2010 Lib Dem vote, the “now voting Labour” share has dropped over the last 2 weeks to its lowest level in the last year. Still voting LDs has recovered somewhat but don’t knows are at a record high. Those who left the party for Labour are not too fond of their new home.

  6. CROSSBAT11
    `Is Harris the new Gold Standard of polls, I wonder?`

    A deliberate ploy to alter the poll of polls to favour the Tories I suspect :)

    Regarding Youguv,approval rating towards the lower end compared to -19 yesterday,maybe a slightly pro-Labour sample.
    Also one of the lower Tory leads amongst the over 60`s(42 to 32),maybe due to the NHS factor as was discussed here yesterday

  7. @Crossbat11

    On PB, the Platinum Standard even, as from tonight. They’re a predictable bunch.

  8. @ JimJam

    “Don’t see any obvious reason why the lab lead would widen.”

    Maybe, but the recent IPSOS/Mori and Opinium polls suggest a widening Labour lead and we have had three 4% Labour leads in YouGov polls in the last three weeks. The peculiarity has been the intermittent 1% Tory leads that pop up to queer the pitch but we are starting to see more 3%+ Labour leads appear now in a range of different polls. My hunch is that this is probably down to the fading effect of Cameron’s EU veto bounce and, despite some more positive economic indicators emerging of late, rising fuel prices and the NHS Reforms debacle.

    However, as you say, we need to see more polls to fully determine whether my hunch has more scientific substance!

  9. Colin Green

    Interesting change in the tables. From the 2010 Lib Dem vote […] don’t knows are at a record high.

    Not tonight they’re not. 22% as opposed to 26% in the Sunday Times (total 2010 LD non-voters 24% as opposed to 33%). Tonight’s figures are here:

    http://cdn.yougov.com/cumulus_uploads/document/c2z51fxho8/YG-Archives-Pol-Sun-results-050312.pdf

    That said the 2010 Lib Dem non-voters have been high for a while , so tonight looks a bit of an outlier.

  10. Neigh,it was the horse that moved the polls.

  11. You must read The Green benches today and see this video of Lanley’s visit to the Royal Free today. A patient reports that the BBC were there but they chose not to report on the visit. No wonder Labour is struggling to get ahead in the polls with a media blackout on news.

    http://eoin-clarke.blogspot.com/2012/03/wow-leaked-video-footage-of-lansley.html

  12. More show and tell from partisan liz…. I’ll take “must read” with a pinch of salt.

  13. Something else which I think is having a larger effect than it should is the size of the 18-24 vote sample. There’s normally about 90 such respondents but they’re getting weighted up to 209, or about 12% of the total poll. There has been the occasional Conservative lead in this category, some neck and necks, and some 25pt Labour leads.

    Spread out over the second half of February this group ended up at about Lab 50%, Con 33%, but there were some big outliers along the way.

  14. @ALAN

    Well if it was Labour in power then you would be accusing the BBC of being biased in not reporting this.

  15. LIZH
    `No wonder Labour is struggling to get ahead in the polls with a media blackout on news.`

    Totally agree with you Liz
    The video is shocking.You would have thought any media company would have killed for footage like this to increase viewership figures and yet none of them showed it.

    After the workfare debacle,the BBC has been warned off

  16. The Harris poll interests me, if only for it’s sheer weirdness… treating them as normal topline figures with others plugging the gap just makes no sense: Labour are simply not going to be on 23%, and others are not going to be anywhere near that high.

    If they’ve just failed to filter out don’t/knows/won’t votes then that’s more plausible, but it would still produce surprisingly good scores for the Tories and the Lib Dems, and an unexpectedly low mark for Labour.

    My guess is that it’s a typo and Labour are actually on 32%. :p

  17. Agree with Jim Jam’s sensible analysis.

  18. @LizH

    Thanks. I wonder how those four nurses in Newcastle under threat of dismissal for much the same are getting on? It’s also on YouTube so it might yet get a bit of wider coverage.

    http://www.youtube.com/watch?v=p0lPJFEjMKo

  19. LIZ

    I wouldn’t use these forums as my sounding board for every tidbit/political campaign I wish to support.

  20. LIZH……….I doubt if the BBC have blacked out the news, the woman in the film was the Union activist that accosted Lansley in Whitehall a few weeks ago, accompanied by the usual assortment of Trots and troublemakers, nothing worth reporting on. :-)

  21. The BBC is now so far stuffed into the Government’s pocket it might as well be News International. Thompson did his job and now Patten is going to dismamtle them.

    Nobody voted for the privatisation of the NHS, the police, Royal Mail or the dismantling of the BBC. And nobody is reporting it on telly except Channel 4.

    Shameful.

  22. Ken

    I didn’t actually go to check the video yet, if it’s only more images of THAT old harridan, thanks for saving me 5 minutes of my life ;)

  23. Back to tonight’s poll. In terms of 2010 voters, there are too few Conservatives, just as there were too many in the ST poll. It still feels that YouGov is at around 2%, slightly at the low end of the range of the other polling companies.

  24. This could be a blip, Conservatives have brought themselves back into contention over the last few months… it could be that Cameron is going some way to gain the trust of the electorate.

    Gary Gibbon reports that both Cameron and Miiband set store by the mayoral election… a Boris win points to an OM in 2015 for the PM, whereas a loss for Ken will raise fresh doubts about Ed’s leadership (a London-centric view, but that is where so much of the narrative is generated).

    *If* we there is a widening of the Labour lead though, we might be seeing something of a repeat of the situation from Sept 2010 onwards. Of the first dozen polls showing a Labour lead only three or four of them were YouGovs… which is paradoxical given how frequently they report – is it possible that for some reason YouGov is slower at picking up the trend?

  25. Of *far* more interest/ utility for this sight than anything on the blinking TGB (!!) is the latest EC projection.

    *****

    There has been an update to Electoral Calculus published on 5 March 2012 at

    http://www.electoralcalculus.co.uk/

    February saw gains for the Labour party at the expense of the Conservatives. All but one pollster saw Labour gain ground, with an average increase in the Labour lead of around 3%.

    The most recent polls from the five pollsters who published polls in February are:

    ICM (Guardian) has Con 36, Lab 37, Lib 14
    Populus (Times) has Con 37, Lab 39, Lib 11
    ComRes (Independent) has Con 37, Lab 40, Lib 13
    Ipsos-MORI (Reuters) has Con 35, Lab 41, Lib 12
    YouGov (Sun) has Con 36, Lab 41, Lib 11

    Overall the average is Con 36 (-2), Lab 40 (+2), Lib 12 (+1).

    The new national prediction is that Labour will have a majority of 14 seats,
    winning 322 seats (up 20 seats since 29 January).

  26. oh bloody iPhone autocorrect (and bloody lack of ‘edit’ function)…

    “sight” = “site”

  27. ALAN……………In my search for truth, no sacrifice is too great…! :-)

  28. @PHIL

    You would think we were living in a dictatorship instead of a democratic country. No government regardless of their political allegiance should be allowed to have a hold over the media in this way.

  29. BILLY BOB
    ` it could be that Cameron is going some way to gain the trust of the electorate.`

    His leadership figures have returned to pre-veto levels though

  30. SMUKESH – “Also one of the lower Tory leads amongst the over 60`s”

    … don’t go there. Taking one of the sub-sections of a poll can be highly misleading.

    Anthony Wells can feel free to correct me, but it’s always more sensible to take the topline figures, and to avoid drilling down into the figures in one isolated poll.

  31. @KEN
    “the woman in the film was the Union activist that accosted Lansley in Whitehall ”

    You are being economical with the truth. The main person on that video is a doctor with over 30 years experience in the NHS.

  32. MIKE HARTLEY
    `… don’t go there. Taking one of the sub-sections of a poll can be highly misleading.

    It`s the continuation of a recent discussion when in the recent Ipsos-MORI poll,Tory lead amongst over 65`s had reduced considerably as pointed out by BILLY BOB,and I am just looking out for a pattern in other polls

    `

  33. AW – If YouGov had a real outlier of a poll, a la Harris, would you be contractually bound to publish it? Or could you refuse?

  34. And I don’t mean a 6% Lab lead, although that’s also an outlier :)

  35. @SMukesh

    IpsosMORI had a Tory lead among the 55-64, a tie among the 65+, and a sizeable Lab lead among the 75+.

    Only one small sample though. ;)

  36. RAF

    How do you determine if a poll is an outlier due to sample variation? (Other than going back and doing the same poll with a MUCH bigger sample size)

    I think attaching the normal caveats should be sufficient.

    I suspect something is very weird with that Harris poll but if they haven’t screwed up the maths, then is should be published.

    Not publishing something because you don’t like the answers is essentially doctoring your results.

    I don’t think anyone is taking Harris too seriously yet, so publishing a result with very weird numbers doesn’t do anything bad.

  37. LIZH…………I was right………! :-)

  38. @Raf, @Alan

    Can you or anyone recall a single instance of a polling company ever admitting an error in any published poll, and republishing corrected figures?

    Surely mistakes must on very rare occasions be made. A 0% error rate is about as believable as a 100% vote for Putin would have been, even in this age where database programmes reduce the potential for error. It might have something to do with the need to avoid claims for refunds from clients.

  39. Which territory are those Metro/Harris numbers supposed to be for? If they really are meant for the UK, even if they included the “Don’t Knows” and “won’t vote” as parts of the “100%”, this tastes badly of voodoo.

    If we’re talking about just parts of the UK, I’d spontaneously think “… any weird by-election going on that I’m not aware of?”

  40. Here we go again, another spiky poll out of nowhere. We get them all the time no matter what the lead really is, detected by the rest of the polling at rather more mundane levels.

    As for the Harris one – it seems rather like European Election polling before the 2010 election. Could it be they are not asking about Westminister?

  41. PJ

    While the figures look awfully strange, until the tables are up we have to wait and see the data before working out why they are so strange.

    The term voodoo poll usually refers to self selecting, or otherwise unweighted polls that have no business being grouped in with polling.

    If the Harris poll has gone through proper methodology and weighting, it isn’t a voodoo poll. If the sample size is about 100 then this would be a believable and valid poll, although useless due to the size of the margins of error.

  42. ALAN

    With all due respect,can`t see others getting 33% as indicated by Harris.Unless the polling is conducted in a conservative area in Scotland

  43. It looks from the Metro story

    http://www.harrisinteractive.com/vault/HI_UK_News_Metropoll_UK_05-03-12.pdf

    that they have published the % of all respondents, not just those intending to vote.

  44. Smukesh

    I’ve said I find the figures very strange and should be treated with extreme skepticism.

    However, if the poll was conducted with proper methodology, it isn’t a voodoo poll was all I was saying.

    There’s little point discussing it (or trying to come up with nomenclature for the poll) until the tables are up. Once the tables are up, all will become clear I imagine.

  45. @KEITHP

    “Here we go again, another spiky poll out of nowhere. We get them all the time no matter what the lead really is, detected by the rest of the polling at rather more mundane levels.”

    I’ve been expecting a 4-6% Labour lead popping up for a while now. The veto blip is all but gone, and yet the Con VI has remained in the high 30s, rather than the low to mid 30s.

    Looking at the Scottish Lib Dem numbers, back on the 2nd of February they got 9% (again tonight), but look at the number in between:

    9, 7, 5, 3, 5, 2, 5, 7, 6, 4, 7, 3, 7, 4, 4, 4, 6, 3, 7, 6, 4, 4, 9

    Average of 5.26
    Median Abs Dev of 4.75

    Basically 4,5 and 6 are included with all else being outliers.

  46. @OldNat

    Thanks for that link to Harris/The Metro. Limiting the sample to actual supporters, the Con lead must be in the ball park of a far more plausible 11% to 13% then, rather than a paltry 9%. And it’s likewise impressive that the fieldwork was in 2012 – in fact a mere 11 to 14 days ago.

  47. If I’ve read the tables correctly, this could be a turning point of a poll… cue Roger Mexico, stats in hand to put the case for it being otherwise. ;-)

    Look at the tables: A 7% switch of 2010 Con voters to Labour with only 2% from Lab to Con. So, not the bobbling around of don’t know/ others etc.; a net 5% directly from Con to Labour. It’s been a while since we saw anything like that.

  48. @Amber Star

    Be careful. I’m worried about you collapsing in mid cartwheel when Tuesday’s YouGov turns out to be back to level pegging.

  49. StatGeek

    7 outliers in 24 polls? Seems unlikely to me.

  50. @ Amber Star

    “If I’ve read the tables correctly, this could be a turning point of a poll… cue Roger Mexico, stats in hand to put the case for it being otherwise.

    Look at the tables: A 7% switch of 2010 Con voters to Labour with only 2% from Lab to Con. So, not the bobbling around of don’t know/ others etc.; a net 5% directly from Con to Labour. It’s been a while since we saw anything like that.”

    I wonder how what would translate to seats and if that would give Labour a majority in an election. I feel your excitement and cartwheel. I saw a FOX News poll today that made me almost flip cartwheels.

1 2 3