According to the Press Association, YouGov’s daily poll for the Sunday Times meanwhile has topline figures of CON 38%(nc), LAB 33%(+1), LDEM 17%(nc). Once again, there is no significant change and no sign of the Conservative recovery that ICM picked up. Other than the 2 point lead last weekend which in hindsight appears to have been a blip, YouGov have now been showing a lead of around 6 points for a fortnight, with no obvious effect from either the Brown bullying allegations or the Ashcroft tax arguments.

More tomorrow when we see the full Sunday Times results.


131 Responses to “YouGov daily poll 38/33/17”

1 2 3
  1. I suggest a closer reading of the article. And particularly the nice big graph they put up…

    Yes, in the 10 point marginals, the Tories have a 7 point lead. *But*, in the much more important 15 point marginals, they only have a 1 point lead! They’re actually *under-performing* in the target marginals needed to put them over the top.

    So the ‘out performing in the marginals’ meme was sort-of correct according to BPIX. The problem being that they’re not out-performing in all the marginals they need. Instead, out-performing by a huge amount in a small number of marginals they were likely to get anyway, and under-performing everywhere else.

    That’s pretty bad news for the Conservatives, and I can see why the Daily Mail buried this result in an unrelated story.

  2. Rob, You support Lab – we knew that.
    I give informed honest and balanced opinions. My posts have been credible and consistant.

  3. Jay, I think that’s a slight misreading.

    You wouldn’t expect the Con lead in Labour held seats to be as high as their national lead (a bit like similar polls for Wales and London).

    The fact there the marginal difference is only 1% doesn’t mean they are underperforming, as it is a mistake to compare it with their 2% lead overall. In fact the poll is telling us that compared to UNS, even in that second group of less marginal seats the Tories gain one more seat. Small potatoes, sure. But still outperforming UNS even in those seats.

  4. Yes dwin

    is that why you posted this on the other thread?

    ”Love the Lab back slapping and spin, and once again ignoring the obvious – the tory lead increasing.
    Does make me laugh. GB in Afganistan – simply incredible defies belief.”

    March 7th, 2010 at 12:45 am

  5. In a sense, guys (and I mean the left-leaning guys, and gals) you are focussing on the wrong part of the poll. The marginal figures are actually the more comforting part of this poll for Tories. The killer is the 2% national lead. That’s awful. A more sensible approach to this for you would be “yeah, so there is a slightly better showing in the marginals for you Tories than overall, but what does that matter when your lead has almost disappeared”.

  6. Rob, before CH4 poll i accurately informed you against your assertion the the Cons were doing better in marginals. I also told you (to balance your view) that a 39% Cons polling nationally + better in marginals would see a Cons victory.

  7. dwin

    ”No surprise, but some regulars are missing the big picture, the economy, GB, and more importantly the polling data over the last few weeks. Tories are ahead, are on course for most seats, doing better in marginals etc, etc. Smell the coffee Please.”

    The bigger picture is the gap closing fast. the marginals will not deliver the seats needed for a Cameron government. There is not point stating the Tory lead as if that will be enough. It has to be big enough for it to matter. Dwin you need to see the details to realise if the trend continues labour may win the election in May.

  8. @Neil A

    D’oh. You’re quite right. I’m half asleep and got a tad jumbled up there.

    But yes, it does show that the “out-performance” isn’t big enough to get the tories their target seats with the low leads seen in recent polling. They’ll pick up a lot of the “New Labour” gained seats, but not make progress beyond that as the lead differential drops off sharply.

  9. Steve A, I genuinely have no political alliance. I despise them all and wish all government departments were run by experts in that field. I do however have pride in the UK and do study political data. My only aim is to balance opinion and inform.
    Someone said the other day Cons ahead of SNP in Scotland – They were also corrected.

  10. @DWIN

    “Rob, before CH4 poll i accurately informed you against your assertion the the Cons were doing better in marginals. I also told you (to balance your view) that a 39% Cons polling nationally + better in marginals would see a Cons victory.”

    …and I am “accurately informing” you here (as I did after C4 marginals poll) that the *current* differential in the marginals (called amusingly in MoS ‘the Ashcroft effect’) is not enough to get the Consrvatives a majority of even one seat irrespective of the National figures !

  11. re the above discussion about whether Yougov use the same people or not, do they use a pool of undecideds only (of which they select a random subsample and ask what they think today)? Perhaps thinking that it might not be worth asking definates every day because you’ll almost certainly get the same answer over and over (and that would make a tracker poll pointless).

    Or do they really go out and find a whole fresh batch of people to ask every day? That doesn’t seem too easy.

  12. Neil A

    “In a sense, guys (and I mean the left-leaning guys, and gals) you are focussing on the wrong part of the poll. The marginal figures are actually the more comforting part of this poll for Tories. The killer is the 2% national lead.”

    Respectfully diasgree:

    1) The lead is most probably not 2% (just like its most probably not 8% or 9% a la Harris and ICM)- so guys and gals of your persuasion have cause most probably to wipe your fevered brows on this element;

    BUT;

    2) The marginals numbers here don’t get the Tories a majority (just like the C4 numbers did not a few days ago);

    3) Even worse for the Conservatives in target seats 92 thru 138 there is no (repeat NO) marginal swing difference compared to national swing AT ALL (!)- as reported in BPIX MoS article.

    Implict in all what you say (here and generally) is the “Labour overstatement thesis”.

    You think that all these polls (whether national or marginal) are giving the Conservatives a lower figure than in reality they should get if the poll were actually an election.

    Well we shall see if themainstream polsters cannot all do better than they did in 2005 when the ‘overstatement of labours predicted vote share’ in eve-of-election polls was within the margin of error aside from NOP who were smack on.

  13. @DWIN

    “Rob, You support Lab – we knew that.
    I give informed honest and balanced opinions. My posts have been credible and consistant.”

    You are trying just a tad too hard…..

    LOL

  14. @Andy

    I know the boundary changes help the Tories rather than Labour. I was making a tongue in cheek comment in reference to other other comments I’ve seen which made that ‘labour boundary change conspiracy theory’ in face of the evidence. :D

  15. @Jay Blanc

    “I suggest a closer reading of the article. And particularly the nice big graph they put up…”

    The key part of the article is where they report the data showing that in the Conservatives marginal targets 92 thru 138 there is NO extra swing over and above the national swing reported in the poll as a whole.

  16. neil A

    “For me the more interesting elements are that the “fabled marginals effect” brought up by so many posters is – for the second time in a week (in a non AR poll)- proving to be very much in fact a damp squib. No mention whatsoever of the marginal swing being bigger than the UNS swing. Now you may have said that on a previous thread (in which case let me know the thread and time) but at 1.30 am I don’t propose to go searching for it.”

    I have dealt with this dancing on a pin of yours comprehensively now in several posts in this thread and they won’t need much searching you can just scroll up the page a little.

    You are not going to convince anyone (other than histrionic tories as you say) that my words mean somnething they did not.

    If you want to be seen on here a fair and balanced (non histrionic tory) debater then:

    * stop misrepresenting my points;
    * stop de-contextualisng my words so that they suit your wider narrative;
    * and- on occasion- stop directly misquoting me as you have done on a previous thread.

    I’ll keep a closer eye on your comments from now on !!

  17. Rob, fact is i informed you (against your insistance) before mid week polls – that Cons were doing better in marginals. People can read both our postings and work out who was right.

    Neil A and others post very interesting posts, not always in favour of Cons which i have made clear i tend to agree with – he does not always agree with my posts along with a majority of readers – which is absolutely fine. It is your continual spin of what you have stated previously which we all object to.

  18. @ DWIN & NEIL A

    I agree with ROB – you are splitting hairs.

    If the swing in the marginals doesn’t get Cameron into No. 10, then I could not care less whether it is a fraction over UNS.

    Sadly, I also have to respect ROB’s assessment that the CON lead is more than 2%.

    You should also consider this part of ROB’s comment before trying to convince us that his analysis is biased in LAB’s favour.

  19. Steve A , stand by my comment about GB in Afgan.
    If you read the posts on the other poll results tonight, they were very heavily labour bias, and were ignoring the majority of recent polls and focusing on one or two not so bad ones. My comments were honest – sorry.

  20. Why are Yougov being treated as the bellweather polls? Their sampling is biased. their polls show considerable bias to Labour! After the election, The pollsters should be anylised and rated! The least accurate should go out of business. This would stop politisisation of the polling process! YouGov is what it says ProGov. Also can we have a break for a while to let things settle. The polls cause trend and waves, they should be weekl;y not daily!

  21. WMA remains 38:32:17 – it would indeed be interesting to see some more non-YouGov polls. But actually their average deviation from the WMA is 0.0 so there is no evidence that on average, over the long run, they over- or under-claim the CLead.

    But if we only consider the last 2 months, YouGov has on average been understating the CLead by about 1.4, compared to the WMA. And it may be that the real YouGov anti-C bias is nearer 2 points since there are so many more YouGov polls than the rest. I suspect that the assumptions behind their false-recall waiting may be breaking down. Will be interesting to see come the election. And another reason for using a WMA of all the pollsters.

  22. Well if we include the BPIX poll the WMA is 38:33:18 and the CLead is shrinking again. I’m a bit sceptical, there is something pretty odd about the figures. According to the poll:
    * Nationally the CLead is only 2%
    * In marginals where the CLead is -0 to -10% they have a lead of 7 points. So this is roughly a 6% swing to the Conservatives.

    At the 2005 election Labour polled 37% and C 34% so on a 6% swing they should be C:40 Lab:31. This is in line with the WMA but not at all with the claimed national vote. Is it possible that someone innumerate at the Mail has added together the National, 10% Marginal and 15% Margainal figures? This would obviously under-estimate the CLead.

    Let’s see what the subsequent polls bring.

  23. POLL ALERT – CELLO mruk/Sunday Times Scotland

    But no Voting Intention figures have been published by the Sunday Times. I wonder why?

    http://www.timesonline.co.uk/tol/news/uk/scotland/article7052648.ece

    http://www.mruk.co.uk/

    According to the BPC list of Company Representatives, mruk is a member, so I expect that we will have the full datasheets published by Tuesday at the latest.

  24. 2 comments about the BPIX marginal poll.
    13 seats is significant and could make the difference -remember lower seats on the list go anyhow but with a bigger majority, so the lift would be around 13 even if only 50 are taken.
    Key fact for me is that under targae 75 Labour had more or less given up hence the bigger Ashcroft effect.

    The smaller marginal swing premium in the next group of seats is predictable as con tagets 75-150 is where Labout has focused its’ lesser resources.
    These plus Con targets from the LDs (where the incumbent LD will do better than UNS) is where the GE wil be decided.

  25. Neil A – “I think its unlikely to be a “rogue” (unless some methodological problem is found) ”

    {Puts on special pedant’s hat} A poll that had a freaky result because of a methodological problem would not be a rogue. A rogue poll describes the one in 5 polls that are outside the normal margin of error due to the the normal operation of random sample error. They are unavoidable and are equally likely to effect pollsters with excellent or flawed methods {takes off pedant’s hat}

    Keith – new people each day. There is a hard block on anyone being invited more than one every fortnight or so, but even beyond that drawing a sample of 1400 from a panel of 300,000 means it is unlikely anyone will be invited too often.

    Asking undecideds only is a good idea in principle…but it practice it wouldn’t work at all. People are not that hot at saying how likely they are too change. In the past I’ve resurveyed people who said they were absolutely certain to vote for a party a couple of months later and found that some of them have in fact changed their voting intention. Obviously people who say they might change are more likely to, but some who say they won’t nevertheless do.

  26. So with another poll showing a 2% lead it validates last Sunday’s poll which last week looked like an outlier. 2% still remains the bottom of the spread but now looks more solid than ICM’s 9%.

    Still no sign that the Conservatives are getting anywhere near a majority.

  27. “……sampling is biased. their polls show considerable bias to Labour…………politisisation of the polling process………YouGov…..is…. ProGov. ”
    I have my own theory of their methods. You don’t really really think they speak to thousands of people every day and somehow use mathematics to predict the results? Ridiculous. Clearly the pollsters are using the black arts to their own ends. Anthony and his friends have sold their souls, they sit in caverns in hooded capes and chant spells until the results are then revealed to them in goblets of blood or something.
    So think again before libelling them. You really don’t want to get on the wrong side of the Dark One.

  28. Neil A has been posting on the site for a long time and despite our declared different political outlooks we often agree on what the polls suggest.
    Some fellow Labour supporters on here should, in my view, try to be more balanced in their analysis.
    For balance can those You Guv bashing Tories stop.
    Go through the archives and you will see that You Gov and ICM have long been considered the most credible by posters across the spectrum.
    In any event, as others have said, the direction is the most important thing, we don’t know if YG, ICM or AR are the most accurate.
    What Neil A has said early in the thread (an example of his balance) to paraphrase is that the Con lead has narrowed mainly due to Labour gaining from others and LDs but also some Tory drop. All polling companies are showing this.
    Also there is clearly a lift in marginals for the cons which could be 1% or 4% and seems to drop in once the first 75 seats have fallen and the advantage may well diminish as the overall lead gets smaller. This ‘maginals lift’ may or may not prove significant in terms of the final GE victor but will be important for MPs/PPCs in affected seats and seat count of course.
    I think the marginal lift could be just enough to see a small con majority (20-35) rather than UNS making them 10-25 seats short.

  29. @Anthony,

    Thank you for clarifying the definition of the word “rogue” in polling circles. Would you think the chances of an individual poll being “rogue” would be reduced by the size of the sample? BPIX’s sample is so big that the MOE will be narrower than average – does that mean that even if it is a rogue then it is probably “less roguish” than other rogues? Just trying to get a feel for pollsters attitude to outliers.

    @Rob,

    You are starting to make me despair. You are one of the (probably the) most regular and lengthy commenters on this site. You must accept that if you are to expose so many thousands of words of your thoughts to the scrutiny of others, then people may want to check them for accuracy and consistency.

    If you choose to believe that you have at no point been saying that there is no marginal effect that benefits the Tories, so be it. All of your previous comments are there in black and white. Others can draw their own conclusions.

  30. “does that mean that even if it is a rogue then it is probably “less roguish” than other rogues?”

    Yep, that’s about the sum of it. 5% of polls with a sample size of 5000 will still be “rogue polls”, but the context of a poll of that size it means the figures being more than 1.4% of the “true” figure.

    Of course, in reality the figures could be much more distant from the real figure because of all the other potential sources of error from pollsters methodologies – it’s just pinning down the potential amount of error from normal sample variation.

  31. @AMBER STAR

    “Neil A DWIN
    I agree with ROB – you are splitting hairs.

    If the swing in the marginals doesn’t get Cameron into No. 10, then I could not care less whether it is a fraction over UNS.

    Sadly, I also have to respect ROB’s assessment that the CON lead is more than 2%.

    You should also consider this part of ROB’s comment before trying to convince us that his analysis is biased in LAB’s favour.”

    Cheers- I’ve always worn my colours on my sleeve and concomitantly always tried to be realistic and factual with a couple of (either over excited or irritated) exceptions.

    I’m not responding to comments anymore that take clauses of my sentences (not even the whole sentence!) totally out of their wider context and message and then split hairs and dance on the head of a pin in order to stretch a dubious point to a fallible conclusion.

    It’s not clever or useful.

1 2 3