Opinion polls are a little light at the moment, and probably will be for the next few weeks. Even at the best of times there is little polling in the weeks immediately following a general election – we’ve just had an actual general election to judge people’s voting behaviour, attention is elsewhere and newspapers will generally have blown their polling budgets in the campaign. I’d expect even less polling over the next few weeks because of the errors in the polls at the general election. Some of the long running trackers like the ICM/Guardian series and MORI political monitor will likely continue just to avoid a gap in the data series, but generally speaking most of the regular polls will probably pause for a bit while they work out what went wrong and sort out solutions to it.

As it is, the next political events we have too look forward to aren’t about Great Britain anyway, but the Scottish, Welsh and London elections next year – I’m sure polling on them will start firing up in the next few months. The other, more immediate, race is the Labour leadership election.

We have had a little polling on that already – the YouGov/Sunday Times poll at the weekend (results here) asked the general public their preferences for Labour leader. Chuka Umunna came first on 17% (fieldwork was conducted before he withdrew), followed by Andy Burnham on 14%, Yvette Cooper on 8%, Tristram Hunt on 3%, Liz Kendall on 2% and Mary Creagh on 1%. Amongst Labour’s own voters Andy Burnham was ahead on 22%, with Chuka Umunna on 19%.

Obviously the key conclusion here isn’t really who is ahead… it’s how low anyone’s figures are. 55% of the general public said don’t know, 40% of Labour voters said don’t know. YouGov also asked separately about if people thought each of the contenders would make a good or bad leader, and in each case a clear majority of respondents said they didn’t know or didn’t know enough about the person to say. This is a race where the public simply aren’t familiar with the personalities of the candidates to have any clear opinion yet. That’s not necessarily a bad thing for the next Labour leader – the public having no clear image of you is better than having negative baggage – it just means they need to be pretty careful to make sure people’s first impressions are good ones, as they are difficult to shift once the public have formed an impression.

On the other outstanding issue – what caused the polling error – I’m beavering away at looking at what caused the errors and how to put them right, as I am sure are the other companies. I’m not planning on giving a running commentary, though I gave some thoughts at the end of last week on Keiran Pedley’s Polling Matter’s podcast here.

463 Responses to “Polling in the coming weeks”

1 2 3 10
  1. First

  2. I took an interesting poll for YouGov last week which asked for my party preferences in order, AV-style. Is that likely to be published?

    First. [Second, Third, etc…]

  3. That was an excellent podcast if I may say so ;) Full lineup:

    John Curtice (U Strathclyde)
    George Foulkes (Labour)
    Damin Lyons-Lowe (Survation)
    Rob Vance as summariser
    and host Keiran Pedley

  4. Surely a big answer to the question of polling error during a campaign, must be to add in questions and information that make the VI Q more like the actual voting decision in the ballot booth.

    I.e. Asking more questions about what people think about the state of the economy, how well/badly people think the respective leaders are performing, who would make the best PM, then finally asking the question of “would you support DCs Cons or ANON’s Lab?”.

    Whenever the names were added into the VI Q like this during the last parlt, there was always a four point (ish) spike towards the Cons. I always said that EMs ratings would drag down the vote relative to the VI published figures (and against those saying his poor ratings were “factored into the price” – they weren’t as the more specific question showed).

    AW – is there any chance companies will begin to post their headline figures as support for the leader+Party??

  5. Mr Attlee was asked why Labour lost and he replied: ‘Not enough votes’.

  6. CL1945,

    Was that Attlee? I remember reading it was what Alf Dubs (Labour MP for Battersea) wrote on his Job Centre form as his cause of unemployment after losing in 1987.

  7. Adrian B – who knows, I’m not pre-empting any conclusions.

    All I’ll say is that we need to be careful not to jump at anything that would have produced more Tory figures and assume that method is more right. Including leader names in the question did indeed increase the Tory lead…but that doesn’t necessarily mean a lack of that factor was the cause.

    What if there was some other reason that meant the samples were too Labour? If that was the case we could change to wording that emphasised the party leaders, and the polls would be fine until we got to a situation when the Labour leader polled better than the Conservative leader… then the polls would be doubly wrong (the original error, and a new leaders in the question error!).

    I’m sure people can think of a hundred different ways that pollsters could have done things that would have produced more Tory answers and, therefore, been more “right”. We need to find a problem that actually addresses whatever the cause was, then hopefully it will keep being right.

  8. @ NC

    It’s nice to see you here too (not only on your lovely blog). Congratulations for that post by the way.

  9. Anthony

    You also have to make sure that you don’t get the English polls “right” and simultaneously screw up Scottish & Welsh polling – which is going to be more relevant over the next year.

  10. Dan Jarvis is backing Andy Burnham for Labour leadership.

  11. The biggest next political event we have to look forward to is the EU referendum on May 5th 2016, the same day as the elections mentioned in the article. Probably.

    Definitely Attlee, and Alf Dubbs quoted him.
    Alf Dubbs gets a very good write-up in O’Farrell’s book about ‘Things can only get better’ which is very funny.

  13. Yvette cooper wants labour not to be anti business and would cut corporation tax(gawd help us ) -looks like she is after the blairite votes from chuka and kendall.

    Turning into a two horse race.

  14. AW

    I know I am amongst several who have mentioned this explanation for the inaccurate before but I haven’t yet seen a conclusive answer :

    It is simply that pollsters failed to completely discount anyone who was not registered to vote. This should have been the first Q and a “No” should have triggered “thank you and goodnight” unless pollsters wanted to alert anyone to the fact that Labour in particular harbours a large number of the unregistered.

    It also made me query canvassing techniques – Labour and I suspect other parties spend a lot of their time cementing and reminding their known, registered supporters. In addition they should have been knocking on all the doors not on the electoral register – and of course well before the registration deadline.

    Individual registration was identified there years ago by a number of Labour activists as a dagger to Labour’s heart. I fear it did make a major difference. If I had not reminded my own student daughter she would have lost her vote. Thousands of students and transient young workers did. Not mainly Conservatives I think.

    Hope you’ll look at this anyway.

  15. Ms Cooper sounds as if she has learned something from recent setbacks. It is just her taste in men that is so suspect.

  16. Started listening to that podcast – starting with the assumption that they got the Ukip result right is bogus imo. It might be true but there’s an equally plausible explanation that the Ukip polling being wrong is part of the cause of the whole polling being wrong.

    If there was a late swing to Con it would be bluekippers most likely did it and the idea that both blue and red kippers would go back equally to the party they came from was never likely imo.

    Using that premise instead would mean the polls under-estimated the Ukip vote all along but the late blukip->con swing partially disguised it.

    To broadly test this idea you could take the Ukip numbers from the early polls and then instead of redistributing a percentage of Ukip votes back to their previous parties see how it looks when you do it with just the Con kippers and not the rest.


    Another grooming gang trial – Aylesbury this time. If the media reported the full scale of it over the last 16 years then the map of the worst areas would align almost perfectly with the Ukip peak vote.

  17. @Chrislane, Frank talking!
    Reminds me of my grandad, he used to read the obituaries and always scoffed at cause of death. There is only one cause of death, he used to say, ‘shortage of breath’.

  18. @AnthonyWells

    I’m concerned that all the discussions you’re holding are pollsters talking to other pollsters and getting all pollstery and echochambery.

    May I suggest you consider contributions from other branches of the statistical family? I’m thinking of things like Dual System Estimation[1], whereby missing respondents are deduced by linking to a different survey done on a different basis? Happy to explain if you need it.

    [1] http://www.ons.gov.uk/ons/guide-method/census/census-2001/design-and-conduct/the-one-number-census/methodology/steering-committee/papers-by-year/1999-papers/the-role-of-dual-system-estimation-in-the-2001-census-coverage-surveys-of-the-uk.pdf

  19. 56 snp lobby fodders pick on an 83 year hero of the left by pinching his seat-could turn nasty.

    And they want to have a Mr Burns bar .

    They have only been there a day -have they all sworn allegiance to her maj ?

    Quite remarkable.

  20. Sunreada

    Have you read Dennis Skinner’s autobiography?

    None of the newly elected MPs have taken the oath yet – that happens tomorrow.

    Do keep up.

  21. @ Martyn

    Thank you for the link. It is a joy (ok, there are other joys …) to read it.

    … But I don’t think it it the answer for the troubles of this year. However, it is something for the future (probably puts up the prices). I liked the simplicity of the regression function.

  22. Anthony-you didn’t mention the mooted BPC investigation.

    Is it happening?
    If it does will YouGov take part?

  23. No of course not I read the english sun.

    Will try to do betta senior fiftysixer -which ones gonna do it in gaelic ?

  24. I know others are already available — just completed my 2015 general election results spreadsheet Since I’d already started doing this one before the others were available I decided to finish the job. Includes result for all 3,971 candidates:


  25. @ Sunreada

    I don’t know if he’s the hero of the left but there are many snippets of videos, among them a long speech on the video channel from 2013.

    In general, can we not make it a national issue?

  26. National meant to be nationality

  27. @Laszlo

    If I may press the case for the moment. One thing that has been suggested as a cause of Pollargeddon ’15[1] is differential response: Labour voters being overrepresented/Tory voter underrepresented in the response.

    In 1991 the UK Census was in trouble: differential response (younger men underrepresnted/older women overrepresented) posed a serious problem. Dual System Estimation was introduced and worked well in 2001/2011. Given this, it would not be a good idea to discard the approach quickly

    [1] I so want to be Chris Morris when I grow up…

  28. @AndyJS

    Have you got a blog or other location where your work is collated?

  29. @ Mr Jones

    To assume a massive exodus of Labour to UKIP would assume a massive vote for Labour of ex LibDems who haven’t voted for Labour (because it would have to be bigger than the switch between 20025-2010) before at a similarly massive scale.

    But in any case you just have to accept that is very likely that a very large number of 2010 LibDem voters (with Lab, Con) voted for UKIP.

    A strange transient 3 million voters … But they are entitled to it. Try to hold of them.

  30. A question for overnight – when will we start getting regular polls again? (and mentally upweighting the Tories by three points in each one).

    I’m off to bed – got to be up early so the SNP don’t nick my usual seat in the library.

  31. @ Martyn

    No, I understood that – but I have doubts that polling companies used census data. if they did, then it’s fine even for the past (on the other hand on YouGov NowCast there seems to be some census data … And it was wrong – the forecast) and I’m corrected.

    Sometimes I wish papers from government departments were easily accessible. I have access to Treasury ones (obviously not all) and they are extremely useful. They also do microeconomic and accounting papers too. Thorough and rigorous.

  32. Mr N

    What? As well as the 40 seats of yours they “nicked” in Scotland? :-)

  33. ON,

    A humorous reference to the rather unusual display today with Dennis Skinner!

  34. Mr Nameless

    What is in the library? Students go there only if they get credit for it. :-)

    Good luck, by the way.

  35. @Laszlo

    No, that’s not it. DSE is a technique, not a dataset. It’s a way of linking two samples from the same population together to deduce which respondents were missing. It’s nothing to do with census data per se.

  36. Mr N

    I had realised – though the lack of seating is clearly foolish when the concept of two tier questioning is well established.


  37. Yes rollingauloise -bit rum innit.Ms cooper gone all business friendly.

    Thought it was only your great party and the snp who wanted to cut the company tax .

    Singapore here we come !!

  38. Laszlo

    “But in any case you just have to accept that is very likely that a very large number of 2010 LibDem voters (with Lab, Con) voted for UKIP”

    I don’t need to accept anything as I’ve been saying for years part of the Lib vote in northern areas was proto-Ukip i.e. ex floating or Lab voters who wouldn’t switch to Con like they did in the south (cos Thatch ) so they went Lib instead.

    The people you think are distinct are the same demographic. The only difference with these Lib voters is they already moved away from Lab in in earlier elections and the move to Ukip was the last step.

    For that demographic in the south the shift over time was more Lab->Con->Ukip and in the north more Lab->Lib->Ukip. (cos Thatch)

    What you just have to accept is Lab has lost millions of their old core vote for very straightforward reasons that are only a mystery because the media are reluctant to report them.

    “To assume a massive exodus of Labour to UKIP would assume a massive vote for Labour of ex LibDems who haven’t voted for Labour (because it would have to be bigger than the switch between 20025-2010) before at a similarly massive scale.”

    Nope, you’re ignoring the unwinding of that part of the Lib vote that stemmed from opposition to the Iraq war.

    The result (in the relevant seats) was the *combination* of the two things: proto Ukip Lib votes to Ukip *and* the Iraq war Lib vote back to Lab.

  39. Perhaps someone could explain something. Since the bench where Dennis Skinner normally sits was traditionally reserved for the third party, why does he sit there and why is it in some special way ‘his’ since he is in Labour which has never been the third largest party within living memory? Just curious.

  40. @ Martyn

    OK. I’m with you. Apologies. I thought I understood the paper and skimmed through the rest concentrating on the outcome.

    It’s really good. The error is still there, but its size is down to the effort. Did I get it right?

  41. “no clear image of you is better than having negative baggage”
    One danger for candidates for the Labour leadership is that some ‘negative baggage’ (eg Mid Staffs) may well not be raised until after the candidate is elected.

  42. @ Dave

    In the case of one candidate (Burnham) it would be a very positive one as he acted on it …

  43. Quite like what I see of the new SNP MPs. A breath of fresh air, hope they shake things up a bit.

  44. @Laszlo

    The technique enables you to estimate the number of people who are missing from your survey and the characteristics of those people. Census uses it to estimate the number of people who didn’t fill out a form in a given area and then estimate the age and sex of those people. If YouGov used it then it could work out the people who didn’t respond to a given poll and then estimate the party those people would vote for.

  45. Voter registration and Labour: when I mentioned the potential problem I was put firmly back in my box and told it wasn’t an issue, though would be after this election. I assumed the agent (etc) were right and in any case was in no position to argue. An impression that different councils acted differently and that ours was ‘OK’?

  46. I have my own idea regarding the failure of the polls. At this election there seemed to be much more in the way of regional swings, swings specific to certain types of seat and other very local effects that I think made the standard opinion poll of 1000 people or so very unpredictable. After all, say a pollster ignores the 200 safest seats in the country, with approximately 1000 people responding, he’ll get 2 people per seat plus a few left over. With such diverse swings actually occurring, maybe this simply made polling 1000 people a lottery. Maybe polls in Scotland and Wales worked because they’re small enough regions for every seat to count in the survey. Maybe if England had been chopped into 11 regions and each region polled separately and the results combined, though not the pollster’s usual method, perhaps it would have given a better result. Maybe that’s why the Comres poll of the South-West got it right.

  47. @ Martyn

    Yes, and they translate the bias to error (although it is not Bayesian, but kind of using the logic of it, but it’s not important). So it would be a way to use design to address the sample. It assumes though a very solid database (actually two).

  48. @ Guymonde

    I don’t know, but Liverpool council certainly put more effort in the registration this year than any other time. It is not a major difference, but it was discernible. Probably they pushed harder in the council estates, but I don’t know.

    But turnout was significantly higher all over Liverpool even though there was no chance that any of the seats would be in danger. And the extra votes went to Labour.

  49. Sunreada – “bit rum innit. Ms cooper gone all business friendly.”

    Wasn’t she part of Alistair Darling’s Treasury when they cut capital gains tax to 18%?

    I think she may have always been business friendly and was biting her tongue hard during the Miliband years…

  50. Ranked share of the vote for Con, Lab, LD, UKIP, Greens, SNP, PC. (Click on sheets at bottom to toggle):


1 2 3 10