There have been several new polls with voting intention figures since the weekend, though all so far have been conducted before the government’s defeat on their Brexit plan.

ComRes/Express (14th-15th) – CON 37%(nc), LAB 39%(nc), LDEM 8%(-1), UKIP 7%(+1)
YouGov/Times (13th-14th)- CON 39%(-2), LAB 34%(-1), LDEM 11%(nc), UKIP 6%(+2)
Kantar (10th-14th) – CON 35%(-3), LAB 38%(nc), LDEM 9%(nc), UKIP 6%(+2)

Looking across the polls as a whole Conservative support appears to be dropping a little, though polls are still ultimately showing Labour and Conservative very close together in terms of voting intention. As ever there are some differences between companies – YouGov are still showing a small but consistent Tory lead, the most recent polls from BMG, Opinium and MORI had a tie (though Opinium and MORI haven’t released any 2019 polls yet), Kantar, ComRes and Suration all showed a small Labour lead in their most last polls.

Several people have asked me about the reasons for the difference between polling companies figures. There isn’t an easy answer – there rarely is. The reality is that all polling companies want to be right and want to be accurate, so if there were easy explanations for the differences and it was easy to know what the right choices were, they would all rapidly come into line!

There are two real elements that are responsible for house effects between pollsters. The first is the things they do to the voting intention data after it is collected and weighted – primarily that is how do they account for turnout (to what extent do they weight down or filter out people who are unlikely to vote), and what to do they with people who say they don’t know how they’ll vote (do they ignore them, or use squeeze questions or inference to try and estimate how they might end up voting). The good thing about these sort of differences is that they are easily quantifiable – you can look up the polling tables, compare the figures with turnout weighting and without, and see exactly the impact they have.

At the time of the 2017 election these adjustments were responsible for a lot of the difference between polling companies. Some polls were using turnout models that really transformed their topline figures. However, those sort of models also largely turned out to be wrong in 2017, so polling companies are now using much lighter touch turnout models, and little in the way of reallocating don’t knows. There are a few unusual cases (for example, I think ComRes still reallocate don’t knows, which helps Labour at present, but most companies do not. BMG no longer do any weighting or filtering by likelihood to vote, an adjustment which for other companies tends to reduce Labour support by a point or two). These small differences are not, by themselves, enough to explain the differences between polls.

The other big differences between polls are their samples and the weights and quotas they use to make them representative. It is far, far more difficult to quantify the impact of these differences (indeed, without access to raw samples it’s pretty much impossible). Under BPC rules polling companies are supposed to be transparent about what they weight their samples by and to what targets, so we can tell what the differences are, but we can’t with any confidence tell what the impact is.

I believe all the polling companies weight by age, gender and region. Every company except for Ipsos MORI also votes by how people voted at the last election. After that polling companies differ – most vote by EU Ref vote, some companies weight by education (YouGov, Kantar, Survation), some by social class (YouGov, ComRes), income (BMG, Survation), working status (Kantar), level of interest in politics (YouGov), newspaper readership (Ipsos MORI) and so on.

Even if polling companies weight by the same variables, there can be differences. For example, while almost everyone weights by how people voted at the last election, there are differences in the proportion of non-voters they weight to. It makes a difference whether targets are interlocked or not. Companies may use different bands for things like age, education or income weighting. On top of all this, there are questions about when the weighting data is collected, for things like past general election vote and past referendum vote there is a well-known phenomenon of “false recall”, where people do not accurately report how they voted in an election a few years back. Hence weighting by past vote data collected at the time of the election when it was fresh in people’s minds can be very different to weighting by past vote data collected now, at the time of the survey when people may be less accurate.

Given there isn’t presently a huge impact from different approaches to turnout or don’t knows, the difference between polling companies is likely to be down some of these factors which are – fairly evidently – extremely difficult to quantify. All you can really conclude is that the difference is probably down to the different sampling and weighting of the different companies, and that, short of a general election, there is no easy way for either observers (nor pollsters themselves!) to be sure what the right answer is. All I would advise is to avoid the temptation of (a) assuming that the polls you want to be true are correct… that’s just wishful thinking, or (b) assuming that the majority are right. There are plenty of instances (ICM in 1997, or Survation and the YouGov MRP model in 2017), when the odd one out turned out to be the one that was right.


1,834 Responses to “Latest voting intention and the mystery of house effects”

1 2 3 37
  1. Thanks, as always, for providing both this site and the commentaries on polls.

    Most of us have learned an enormous amount about the technicalities of polling. It is very helpful in filtering out a lot of the manifest nonsense promulgated by much of the MSM as well as parties and their ardent supporters.

  2. Thanks Anthony for a great resume for those of us who give a sh!t about polling accuracy.
    As to the Elephant in the polity of how YouGov is used by the right to portray ‘that which is not’, ie a consistent Tory lead, then clearly, we’re all on our own in making our minds up!

  3. Obviously YouGov are happy that their sample is representative. I doubt that weighting makes any more difference than adjustments for don’t knows. As Anthony says we won’t know until there’s a test. However I don’t recall such a large systematic difference between companies for a long time, if ever, and I note Anthony doesn’t care to speculate on why YouGov’s sample is “better”.

  4. @Danny

    “So you agree with polling that there is now a remain majority?

    If so, how can anyone who in any way believes in democracy want to push through something opposed by the nation?”

    No, but I believe that there is a clear fed up of this and just get it done majority.

  5. Good Evening all from a wet Bournemouth East, where our Tobias Ellwood MP has spoken against the ‘right’ wing members of the Government; and against No Deal.

    To Deal or Not to Deal, that is the question.

  6. If there was a large proportion of “fed up of brexit” if there was a 2nd referendum surely they would all be voting remain as the quickest way to put an end to it all.

  7. Hireton (if you are around)

    I suspect that you know more about the procedures in the Commons than anyone else on here.

    Have you a view on the potential efficacy of this proposed amendment to allow a motion proposed by 300 or more MPs (including 10 from the Government party), including MPs who were elected as an MP for at least 5 parties, to take control of the business of the House, on a specific day, in order to prevent a No Deal?

    https://www.buzzfeed.com/alexwickham/secret-plan-by-rebel-mps-to-stop-a-no-deal-brexit

  8. I didn’t phrase that well!

    It would be hard to find someone “elected as an MP for at least 5 parties” – though there may be some who have been members of 5 factions!

    The actual wording would allow Independents, who had been elected as a party MP, to be counted – even if they had resigned from or been thrown out of their party subsequently.

  9. A little bit of an addition to Anthony ‘s explanation of the variance across pollsters. The trouble is, as he says, that in absence of raw data, only general claims can be attributed that may or may not be true, or that causal linkages are attributed to random variations.

    https://m.youtube.com/watch?v=ez4DgdurRPg

  10. Thanks for the illuminating write-up Anthony. Unfortunately I have to agree with Oldnat when he says that many of us have learned a lot about polling.

    There are rumours that there may be yet another GE before the end of February, so that would be a good test for which methods are currently more accurate.

    One thing which I haven’t seen discussed is that the importance of various weighting factors will change over time, so that there will never be a ‘perfect’ way of weighting. For instance, weighting by newspaper reading would seem to be less important than it once was, now that circulations are shrinking. Also, I would suggest that weighting by social class would be likely to vary over time – now that there are many fewer traditional working class jobs.

  11. It may be that there is at the moment no ‘right’ answer on who is in the lead. Let’s suppose that as in previous elections much depends on who is going to vote, and as the elderly seem to vote fairly reliably, on whether young people are going to vote. And they seem to fluctuate in their propensity to do this, while those of them who participate in polls do not seem to be typical of their fellows Pollsters, I believe, vary in the way they take account of this unreliable behaviour by the young. However, it is perfectly possible that it is at the moment unpredictable. Corbyn last time round was good with the young and reached parts of the electorate that others could not. However, his Brexit behaviour or simply the fact that in between elections the younger generation have more urgent things to get on with may mean that their commitment cannot be relied on. I listened the other day to Caroline Lucas and despite being in touching distance of 81 I found her inspiring. If Corbyn wants to keep the young vote he may need to up his game.

  12. Bob Maclennan managed to get elected for 3 parties to the HOC in the space of 4 General Elections!!

    May have been others but I can’t think of any?

  13. ComRes tweet

    If a referendum were held today, the result would be 54% Remain, 46% Leave

    (poll for Mirror/Express)

  14. @ Pete B

    And The Sun has made an enormous loss.

    So is likely to be cutting back on journalists and pages. It could become less of a problem for Labour.

  15. via Britain Elects –

    Westminster voting intention:

    CON: 38% (+1)
    LAB: 37% (-2)
    LDEM: 10% (+2)
    UKIP: 6% (-1)
    GRN: 3% (-)

    via @ComRes, 16 – 17 Jan

  16. Peter Kellner

    Jan 19, is Crossover Day. #brexit Leave majority disappears, even if no-one changes their mind. Teenagers, mainly Remain, turning 18, + older voters, mainly Leave, dying, shrink Leave majority by 1350 a day. From Sunday, demographics alone produce a Remain majority

    I suspect that it’s a tad more complex than that!

  17. @oldnat

    It seems to apply specifically and only to the situation we are now in i.e the Government being required to make a statement following defeat on its motion to agree a EU withdrawal deal.

    It would seem to be designed to avoid any further doubt or debate about whether a government business motion is debatable or amendable and does so by giving non government MPs the ability to determine the terms of the motion to be debated as the first item of business. . It requires a double majority: a majority of MPs not on the governmeny payroll ( ie 300 out of 500 assumimg the payroll vote at about 150) and a majority of parties in the house ( including at least 10 from the government party). Any motion tabled and debated would be amendable.

    I am not clear about the implications of the reference to SO 41(a) but it looks as though that could be intended to ensure that the Government cannot use parliamentary procedure to prevent the motion being debated and voted on.

    Overall it looks a carefully constructed way for a majority of cross party backbench MPs the House of Comons to be in control to make clear that a no deal outcome is unacceptable but whether it would enable the House to indicate an alternative would presumably depend on whether 300 MPs can agree on that!

    The Clerk of the House may have a different view!

  18. Hireton

    Thanks.

    I had wondered about the selection of 300 as the magic number, but your explanation seems right. It’s 50% of the non-payroll vote – itself not an exact number, but 11 MPs don’t/can’t vote anyway.

    In realty, therefore, it’s more than 50% of non-government MPs. I’m sure there has also been considerable head counting as to how MPs might vote, before that comfortably principled number was settled on.
    See my comments previously on the need to separate politician’s statements of principle to disguise expediency!

  19. A brief on the Comres poll below, key to me is 54% would vote remain in a referendum yet only 33% want to remain in the EU. Explain that!

    Voting Intention

    (nb adds up to 98% due to rounding)

    (% in brackets relate to ComRes/Daily Express poll on 16 January 2019)

    Political findings:

    By a ratio of almost two to one, voters support withholding the salaries of MPs and Government Ministers until they have reached an agreement on Brexit (48% to 29%).

    Most British adults “don’t really care how we do it but just want Brexit to be sorted” (52%).

    If a referendum were held today, the result would be 54% Remain, 46% Leave

    Three quarters agree that the Government have handled Brexit negotiations badly (74%) in comparison with three in five who agree the EU has handled negotiations badly (61%). In fact, six in ten 2017 Tory voters (63%) think that their own Government have handled Brexit negotiations badly.

    However, things are little better for Labour – only one in five think Jeremy Corbyn would have done a better job than Theresa May at negotiating Brexit (22%).

    Similarly, fewer than half of 2017 Labour voters agree Jeremy Corbyn would make a better PM to negotiate Brexit (46%), although current Labour voters are a bit more likely to agree with this statement (57%).

    Four in ten adults think Jeremy Corbyn was right to call a vote of no confidence to try and force a general election (38%).

    A clear majority of 2017 voters for all the main parties agree that “today’s mainstream parties do not offer an appealing choice of who to vote for” (Con 64%, Lab 60% and LD 72%).

    Three in five British adults support May and Corbyn setting aside party differences and work together to devise an agreement on Brexit that can secure the support of most MPs (61%).

    Four Brexit outcomes are within two percentage points of each other: requesting the EU to allow an extension of Article 50 (40% support; 30% oppose), forming a Government of National Unity (39% support; 21% oppose), No deal (38% support; 36% oppose), and a second referendum (38% support; 47% oppose).

    Attracting less support are remaining in the EU (33% support; 45% oppose) and calling a general election (30% support; 50% oppose)

  20. @Hireton/Oldnat

    I think that the significance of the 300 is merely that it is a huge number – the real sting is with the 10 tories, which is enough to take the tories + DUP below the working number to achieve a majority.

    It appears carefully crafted to allow 10 tories to sign without the signed motion itself achieving a majority, but leaving the Speaker little room but to overrule the clerks, because quite plainly it has a reasonable prospect of being passed

    Standing Orders are here https://publications.parliament.uk/pa/cm201719/cmstords/1020/body.html

  21. New Opinium out too:

    https://twitter.com/OpiniumResearch

    Labour have flipped into lead on that one, almost opposite movement to ComRes, but it’s all MoE stuff.

  22. Gaurdian reporting that ministers have discussed referendum options with Lib Dems, as part of the discussions on the deadlock. No 10 briefing that May still opposed, but Gove was one of the ministers in talks with Cable.

    Going to be an interesting few days, with a great deal of strain on the government, I suspect.

  23. TO

    The “5 parties” rule is also a “sting”. It would be unlikely that the Brexiteers could find supporters from MPs elected for that many parties – Con, DUP, Lab – but who else?)

    I do wonder who Grieve et al have in mind for those elected for (but no longer in) a party. Independents could count towards the 300 anyway.

  24. Oldnat

    The problem with guessing what difference the youth vote would make is not that the majority of young people are pro remain, but rather what proportion of young people are pro leave and is that percentage enough to balance the votes of those who voted leave and have passed on.
    Together with knowing what percentage of those who would now be able to vote would actually vote . That statement of kiliners seems as you have indicated a tad over simplified.
    The one thing I would be surprised at is should there be a second referendum that the vote will be anywhere near the first referendum in terms of people voting numbers, it seems to me people in the U.K. are so disillusioned with politicians there could be large swaths of the public washing there hands of the whole thing not because they don’t have a view rather they simply don’t trust politicians to deliver.

  25. tonights polls via Britain elects
    Opinium
    LAB: 40% (+1)
    CON: 37% (-2)
    LDEM: 7% (+1)
    UKIP: 7% (+1)

    via @OpiniumResearch, 16 – 18 Jan
    Chgs. w/ Dec

    Comres
    CON: 38% (+1)
    LAB: 37% (-2)
    LDEM: 10% (+2)
    UKIP: 6% (-1)
    GRN: 3% (-)

    via @ComRes, 16 – 17 Jan

  26. Latest Opinium poll in tomorrow’s Observer has Labour on 40% with a 3% lead over the Tories. Labour 1% up and Tories 2% down from last Opinium poll when the two parties were neck and neck.

    Lots of interesting and slightly contradictory stuff in the poll on Brexit too.

    https://www.theguardian.com/politics/2019/jan/19/uk-voters-would-back-remain-in-eu-over-theresa-may-brexi

  27. Seems Con about 37-39% in polls. Labour showing linked to LiB Dem rating. When Lib Dem are higher Labour are lower and visa versa

  28. Two more polls today and YouGov are still sticking out like the proverbial sore thumb. I take Anthony’s disclaimer that “there are plenty of instances (ICM in 1997, or Survation and the YouGov MRP model in 2017), when the odd one out turned out to be the one that was right”, but it’s starting to stretch things a little when it’s consistently the odd one out in a field of 7 or 8 other reputable pollsters.

    As for the danger of ” assuming that the polls you want to be true are correct…”, this of course works both ways and could be equally applied to Andrew Neil and Isabel Oakeshott.

    Oh, and probably Colin of this parish too!

    :-)

  29. ON – does Stephen Lloyd Count – or Kelvin Hopkins as they both independents? Frank Field gave up Lab Whip IIRC?

  30. Matt126

    ” Labour showing linked to LiB Dem rating.”

    Obviously, the dynamic is different in Scotland – and currently we only have geographic crossbreaks to suggest possible shifts, but it does seem to be the Labour vote that is more fluid and vulnerable.

    Here, any “movement” in VI is between Lab and SNP.

  31. Jim Jam

    By the terms of the Grieve amendment, all of them would count. I had already assumed some existing Lab MPs (like Hoey) could be counted on to form a Brexiteer alliance, so Field and Hopkins wouldn’t add to that.

    Lloyd would count as a Lib Dem, but isn’t he pro-Deal and against No Deal?

  32. Even as Independents as the no longer have to Labour Whip?

  33. Jim Jam

    Yes. The draft amendment says “at least 300 members of the House elected to the House as members of at least five parties and including at least 10 members elected to the House as members of the party in Government”.

    Though, when I look at that again, I wonder if Sylvia Hermon would count?

  34. @ Old Nat

    Looks very undemocratic to me, what if sufficient numbers of the Government were to resign for a short period enabling them to return to the back benches? A fix to avoid a fix.

  35. I cannot understand all the fuss about the Greive amendment, the Government would simply consider it not legally binding, just as May pointed out his last one was after winning the no confidence vote.

  36. Bantams

    I made no claim that it was more or less democratic than the current convention that a government without a majority of its MPs in favour of its policy getting to control the business of the House.

    More likely, I would think, is that the amendment itself would fail because Labour wouldn’t want to end the convention that it as a minority government (itself elected on a very minor share of the vote) could be prevented from exerting total control.

    As with discussions of Euref2, arguments as to the “democracy” of any expedient process seem inappropriate.

  37. MrQueue

    In the case of amendments to the Standing Orders of the House, “legally binding” doesn’t come into it. They are entirely a matter for MPs.

    The amendment doesn’t instruct the Government on anything. All it does is to alter the order of business. Whatever motion(s) were debated and voted on would be a matter for the whole House.

  38. Number Cruncher poll spanning the WA vote

    https://www.ncpolitics.uk/2019/01/voting-intention-did-the-last-week-hurt-the-conservatives.html/

    Our modelling suggests a high probability that Tory support was higher in the days before the vote than after it, although it is harder to be precise about how large or durable any impact will prove to be.

    Con 41% (+3)
    Lab 39% (-1)
    LD 8% (nc)
    UKIP 4% (-1)

    (changes c/f August 18)

  39. “I believe all the polling companies weight by age, gender and region”

    Assume this is “region”, as given, rather than “religion”? Just want to ensure I’m clear on what variables are being used so I can dabble with my own modeling (and there are a couple of typos elsewhere).

    Thanks again for all your work here – I know it’s been a frustrating three years, for pollsters and pundits as much as public!

    Kester

  40. “I believe all the polling companies weight by age, gender and region”

    Assume this is “region”, as given, rather than “religion”? Just want to ensure I’m clear on what variables are being used so I can dabble with my own modeling (and there are a couple of typos elsewhere).

    Thanks again for all your work here – I know it’s been a frustrating three years, for pollsters and pundits as much as public!

    Kester

  41. Turk
    “The one thing I would be surprised at is should there be a second referendum that the vote will be anywhere near the first referendum in terms of people voting numbers, it seems to me people in the U.K. are so disillusioned with politicians there could be large swaths of the public washing there hands of the whole thing not because they don’t have a view rather they simply don’t trust politicians to deliver.”

    It all depends on what the referendum question was, but older voters would be likely to turn out in greater numbers than the young, with the attitude ‘We told you Once!’, whereas youngsters would be more likely to just shrug (IMO). Also, on the difference between new youngsters being able to vote, and oldsters dying, has there been any polling on the views of those who were 16-17 in 2016? And also also, I seem to remember that there is a point somewhere in the 40s age group when Leave starts to outnumber Remain. How many have passed that threshold in the last couple of years?

  42. @OLDNAT, JIMJAM et al

    Isn’t the concept of 300 MPs, 5 parties, 10 govt etc just a cherry-picked playground way of saying that even if we lose the actual vote we can still argue we somehow Really Won?

  43. EOR

    I imagine that those who look forward to a No Deal exit from the EU will say so.

  44. EOR (contd)

    Obviously I can’t speak for those who are drafting this amendment, but if you read it in the Buzzfeed link I provided earlier, you will see that it is constrained to allow only motions relating to No Deal.

    It does not preclude Brexit, it is designed as a HoC procedure to prevent only a No Deal one, should the Government provide no other opportunity to stop such an event.

    Since HoC operates like a playground full of spoilt children anyway, your analogy may be appropriate.

  45. Presumably, with elections coming every two years, there’s a better chance of recall being more accurate.

    Having said that, I bet ‘floating voters’ can remember who they voted for in 1979 and 1997 more readily than in 1987 and 2005.

    Never-ending elections might throw up as much bad recall too, if the arguments and people around them don’t change.

  46. This General Election if called is going to be interesting to say the least. Not one poll I have seen puts Labour above 300 seats but the Tories could get as many as 340. It suggests that it will end in yet another hung parliament with the Tories being the largest party, but maybe only just. Quite how that will resolve things I do not know.

  47. @eor

    The Grieve amendment simply concerns getting a motion for debate and vote by the whole House on the order of business and that can only happen if a majority of backbench MPs from a majority of parties in the House agree to do so.

    It’s interesting to see so many people having a fit of the vapours over a modest proposal for backbenchers to have a modicum of control over what they debate on one issue!

  48. @HIRETON
    It’s interesting to see so many people having a fit of the vapours over a modest proposal for backbenchers to have a modicum of control over what they debate on one issue!

    Agree, it’s almost as if some do not want Parliament to take back control:-)

  49. PeteB,
    “There are rumours that there may be yet another GE before the end of February, so that would be a good test for which methods are currently more accurate.”

    Ah but it wouldnt. Or might not. Or might…
    The last election showed a wholly different result to predictions before it was called. On the other hand, we did get a time series where opinion changed and we could see the polls tracking that change until they converged much closer to the actual result.

    So maybe they were getting it right all the time (though some companies did some tweaking on the way, I think). Or, if their aim was to predict the election outcome as we went along, they failed dismally until a few days before. But while people think polls are predicting an election result, that was a classic example where they probably did what they say in the small print and ask in their questions, ‘if there was an election tomorrow…’, when in fact there was an election after a campaign.

    As to the likelihood of an actual election…One will only be called if it is to the advantage of the conservative party. It will not be called because of lack of confidence in the government by conservative MPs, who are voting against each other on Brexit but who are rock solid in confidence motions.

  50. @Hireton & others -when you have a government carrying out a significant course of action which apparently no longer commands the majority support of the population, in doing so they are refusing to countenance collaboration across party lines and are demonstrably only interested in the unity of their own party, and they are hiding behind a somewhat distorted method of government, then something dramatic must change.

    There are no ‘right’ or ‘wrong’ systems of managing a democracy – just multiple variants which have good and bad points, and are more or less well suited for specific moments in time.

    Our parliamentary system has failed, so on the issue of Brexit, something different is needed.

1 2 3 37