ComRes’s monthly online poll for the Independent on Sunday is out today and has topline figures of CON 42%(nc), LAB 29%(-1), LDEM 7%(nc), UKIP 13%(nc), GRN 3%(nc). There is no significant change in support since last month. For those intrigued by the big difference between pollsters (ICM had only a four point Conservative lead, compared to thirteen points here), remember that polling methods are in a state of flux, with all the companies looking at their methods following the failures in May. Different companies have overhauled their methods to different extents, and even those who have made changes have said they may yet make more. ComRes have introduced a new turnout model based mostly on socio-economic factors like age and social class, and this is why they are producing larger Conservative leads than other polls. Tabs are here.

On other matters, I’ve finally started to update the swingometers on the site. The new version of the simple GB swingometer is now up here, now with added UKIP. Currently it just assumes a static SNP vote, as GB national polls are of so little use in measuring support for a party standing only in Scotland – the updated version of the advanced swingometer allowing for separate figures for Scotland, Wales and England will be along in due course.


172 Responses to “ComRes/Indy on Sunday – CON 42, LAB 29, LD 7, UKIP 13, GRN 3”

1 2 3 4
  1. Colin

    That would support the lowest paid, but still hit lots of Tory voters.

    We are here to discuss the political impact.

  2. HAWTHORN

    It doesn’t need saying does it? If GO tweeks this of course there will be “political impact”.

    If you start from the premise that an £80 bn pa deficit has to be fixed , that welfare reform must be part of the fixing & that WTCs specifically have grown unsustainably -then someone, somewhere is going to feel pain.

    At present the press coverage is all about pain for the lower paid. If DC wants to stick to his One Nation Party of the Workers thing, then GO will have to transfer some of that pain to the better paid..

    Who these people will have voted is now history. Who they will vote for in 5 years depends on factors we cannot now know or even forecast.

    But at present the “political impact” for Cons is a) Adverse & b) an open goal for Labour-so I believe that GO will try to mitigate.

  3. Seems bizarre to be taking about low pay when the present minimum wage is above 90% of the world’s wages. Has anyone calculated how much more the NHS and local government will have to pay for this foolishness?

  4. @Hawthorn:

    Can you provide any evidence for your claim that ICM was more accurate than Comres?

  5. JP

    There is no point in doing statistical analysis (given the unreliability of the data that would be p-hacking at its worst).

    In the run up to May, ICM gave the Tories a modest lead over Labour (apart from their infamous 6/7/15 outlier). Their first poll after the election was C:37, L: 31. Their polling overall suggests a Tory lead of around 6pts since the election.

    Comres gave a slight Tory lead that then immediately jumped to a 12pt lead on 31/5/15 after re-basing. That does not inspire confidence in their polling methodology. Any subsequent alterations are untested against reality.

  6. I am very curious by the assertions that tax credits are bad, and obviously have to go.

    The point of them is that they reduce the marginal tax rate – ie “make work pay”, and this, philosophically speaking, is what all parties are in favour of. Conservatives in particular have been very keen to reduce marginal tax rates, and this is part of IDS philosophy in reforming the benefits system. Free market Economist magazine has been a strong supporter of tax credits.

    Yes it is true that this is “subsidising low wages”, but a free market economist will say that if low wages reflect productivity and the price at which goods can be sold in the market, then they are all that can be paid. If the public interest is that people should work rather than be on benefits, even if their productivity is low, and humanity requires that people are not left to starve, then tax credits would seem to be the optimal solution.

    I really haven’t seen a convincing argument why tax credits are a “bad thing”. I have only seen those who are convinced, for whatever reason, that public expenditure must be cut, and even then no good reason why this is better than eg scrapping Trident. This has to be argued, not assumed.

  7. #Hawthorn

    To clarify you are basing this on not agreeing with the poll results AFTER the election?

    The last Comres poll before the election showed a 1% Conservative lead, the last ICM poll showed a 1% Labour lead.

  8. WOLF

    You need to bear in mind that subsistence living costs are minuscule in many countries in the world.

    COLIN

    I don’t accept the premise. Even if I did I don’t think that a contractionary fiscal policy will work. Even if I did think that a contractionary fiscal policy would work, I would favour increases in income tax.

    Even if I accepted all of that, I do not believe that it is politically possible to do as Osborne proposes. The fact that he favoured winning the last election over his initially favoured policy is evidence of this.

    It is the polling impact that is the issue on this site. A serious attempt to run a surplus between now and 2020 would be political disastrous for the Tories. The only question would be if Labour or UKIP reaped the benefit.

  9. JP

    We understand stats on this site which is why we know not to look at the last poll before an election in isolation. This isn’t PB.com.

    [ :) ]

  10. @John Chang

    Tax credits directly effect productivity. If you have to create a business plan for capital expenditure then the fact you can get folk to do the job on low wages subsidised by the state – makes it that much harder to justify and so we end up with poor productivity.

    Recycling is a case in point, no justification to invest in machinery when we can just employ folk on minimum wage, meanwhile Germany cleans up in the EU recycling business, by making capital investment.

  11. The ICM poll suggests a 5-6% increase in the Conservative lead since the election, given the minimal changes to their methodology.

    This is consistent with Comres given the actual election lead.

  12. JP

    The polls have been re-based since the election. You are comparing apples and oranges, even with the same methodology because pre and post-election polls are weighted using different weighting bases. The post 5th May 2015 polls have weighted base which is based on the actual election result.

    AW did a post at the time if I recall correctly. I don’t think I have misunderstood it.

    The figures need to be compared to the actual result for the time being (and taken with a massive fistful of salt even then). This suggests little change since the election with the caveat that all the polls could be nonsense.

  13. That should have read 7th May of course.

  14. @Hawthorn:

    “We understand stats on this site.”

    Some do. Not sure that you do, since you haven’t actually provided any. Even logic will do, or at least something more than assertion based on personal preference.

    Again, can you provide any evidence for your assertion that ICM was more accurate than Comres?

    I will interested in reading your justification for your evidence more reliable than the polls taken just before the election.

  15. Re re-basing, pollsters frequently re-base their data and this usually produces minimal changes to their results (and I believe 4% is the joint LOWEST lead that ICM has shown since the election).

    Comres have actually made a significant change to their likelihood to vote model.

    ICM’s changes are relatively minimal since the election.

  16. via Britain Elects

    CON: 37% (-)
    LAB: 32% (-)
    UKIP: 15% (+1)
    LDEM: 5% (-1)
    GRN: 4% (-)
    (via Opinium / 13 – 16 Oct)

  17. JP

    The first rule of stats is c**p in, c**p out, no matter how much “clever” p-hacking you do. We do not really know the true position of the parties. All we can say is that there is little change since the election based on the weight of evidence across all pollsters.

    I am not arguing that the Tory lead is 4 points. Even if the data was reliable, the margin of error for each party would be roughly +/- 3pp. It is likely to be 6 points, which is pretty good for the Blues.

    If it makes you happy to think the Tories are 12 points ahead then think that. It is possible that Comres are correct and I don’t want to be a killjoy.

  18. JP

    On re-basing:

    Distortions in the polling model should accumulate as time goes on. Indeed, the polls may well have been quite accurate for some time after the 2010 election, with distortions gradually increasing the further on from the date of poll base.

    Re-basing in this case led to sudden increase in Conservative VI. In other words, the distortion would “re-set”.

    The methodological change in Comres is untested.

    As I said, the weight of evidence is a roughly 6 point lead for the Conservatives, with the caveat that the polls might well be nonsense.

  19. JP/Hawthorn –

    Even if polls were always correct, pollsters would still need to make regular updates to their methods – for example, we all weight by age and gender based on the census or on ONS mid-year estimates – hence when the new census data is published all the pollsters will tweak their targets for age and gender slightly.

    The most important of these in political polling is political weighting, as it makes a real difference to voting intention figures. So before the election, most companies were weighting using recalled 2010 vote. Since the election most companies are weighting using recalled 2015 vote. This would have happened anyway, even if the polls had all been correct – companies weight to the result of the last election, not the one before last.

    This should, as you say, make little difference. If your polls are correct, then re-basing against reality shouldn’t make much difference, as your polls should already have been reflecting reality in the first place. If, on the other hand, polls were some distance from reality, there will be a noticeable difference!

    Put it this way. If I drew a sample now using pre-election quotas and weighted it using pre-election targets, then asked those people how they voted in the 2015 election it would still show Labour and Conservative roughly neck and neck. If I did the same thing and weighted it to the actual 2015 result, it would (obviously) show the Conservatives 7 points ahead.

  20. HAWTHORN

    @” I would favour increases in income tax.”

    I acknowledge your honesty.

    But I couldn’t support the premise that the State balances its books on the income side & never on the expenditure side.

    [email protected]” A serious attempt to run a surplus between now and 2020 would be political disastrous for the Tories.”

    If I understood how this disaster is to unfold then I could comment. But I don’t.

    Of course the current policy is to achieve surplus in & after 19/20-not ” between now and 2020 “

  21. Anybody going to forecast the result of the Canadian election? Another battle over austerity by the sound of it, and another one where the pollsters have the potential to get it wrong.

  22. (A) I can understand why YouGov and other pollsters wish to be able to predict outcomes of elections, but I can’t understand why the public’s actual opinion (adjusted of course by gender, class, age, geography etc to ensure a proper sample) is of no interest. I would have thought that in a democracy the difference between the two figures would be of considerable interest. So why isn’t this published?

    (B) The political implications of tax credit cuts need to be viewed in light of standard psychology (see Kahneman inter alia). People hate having things taken away. It is better by a long chalk never to give someone something than to give it and take it away. They will never forgive you. However if you are going to give someone a negative experience, better to give one big hit, than the torture of death by a thousand cuts. There are lots of people affected here, and I can’t believe this won’t show up in polling once the implications in hard cash appear. However you might as well be hung for a sheep as a lamb – it’s pointless “ameliorating” a cut if you are still cutting.

  23. @Hawthorn:

    I am absolutely not claiming that 12% is more accurate than 4% or stating a preference

    At this stage none of us know which polls are accurate.

    However, you are the one who has stated that one pollster is better than the other apparently simply based on not liking their result.

  24. Opinium (Eng only)

    Con 40% : Lab 33% : UKIP 17% : LD 5% : Grn 4%

  25. JohnChann
    “I really haven’t seen a convincing argument why tax credits are a “bad thing”.”

    [Ahem – you really shouldn’t be looking for one here. To repeat, this isn’t a place to discuss if policies are any good or not – it’s incompatible with non-partisan discussion – AW]

    Whether tax credits or their removal are a good or bad thing politically is another matter.

  26. Sky Sports –

    “BREAKING: World Rugby say referee Craig Joubert’s decision to award Australia last minute penalty against Scotland was incorrect.”

    No wonder he broke the speed record leaving the field at full time!

  27. The lack of an apparent shift from the tax credits changes suggests that one of the following is true-

    (1) There’s significant public support for the changes or at least are indifferent towards them.

    (2) The Corbyn factor is so big that it’s offsetting the impact of unpopular changes.

    Neither would be good news for Labour and either would be good news for the Tories who need to “front-load” their welfare cuts if they’re going to be the largest party in 2020.

    All this is largely irrelevant to 2020, but it’s good news for the Tories in the 2016 elections and for Labour it suggests that they will really have to seize opportunities; just the Tories making unpopular moves won’t be enough for them.

    However, I very much doubt that there will be a Tory majority in 2020, simply because I can’t see how a party can govern under austerity and keep on winning. The last time that happened was the 1930s, when expectations were different and Labour was led by a far-left figure with controversial foreign policy vie… Hmmm, come to think of it, I have no idea what’s going to happen in 2020!

  28. Saying that Hawthorn doesn’t understand stats is a bit rich. If anyone doubts – read back to election times or earlier.

    In any case – the trouble is manifolds, however, Argyris’s notion of single loop and double loop learning is relevant. If the polls problem is some incorrectly in maple enter sampling (e.g. Voting intention), then tweaking is fine. If the underlying assumptions about sampling is correct, but needs to be heavily rearranged, then the current path may be OK. If the underlying assumptions are incorrect (e.g. Radicalisation), and instead of evaluating these assumptions tweaking takes place: mental prison.

    It is testable, by the way, but it is not cheap, and almost certainly would require Bayesian statistics. Furthermore, the whole thing could be a wasted investment if the current (well, the last three years) peters out.

  29. Looks like it might be Rubio vs. Clinton. I wonder who misogynistic racist whites will vote for? A Trump third party run?

  30. From Lord Ashcroft:
    This is the exclusive report tonight from Channel 5 Belize on my health. I can confirm these details are accurate https://t.co/izyO3hi3MM

  31. Opinium (via Number Chrunchr Politics):

    Con 37
    Lab 32
    LD 5

  32. @ Charles

    Oddly, the Canadian elections are very muted in British média. However, for the first time in my life, I put a bet, and it is (not really a surprise) on the incumbent being removed.

    If NDP gets in it would be interesting, because compared to Mulcair, Corbyn is a charismatic leader.

  33. Oops I left out some numbers:

    Con 37
    Lab 32
    LD 5 (-1)
    Ukip 15 (+1)
    Green 4
    SNP 6 (+1)

  34. Charles – “Anybody going to forecast the result of the Canadian election?”

    Canadians are hard to predict at the best of times, but at the moment they are in an oil recession.

    The oil state Alberta changed it’s regional govt in May (at the time we were having our general election) because of the oil crisis. (They went left)

    According to the following article the Albertans are a super tough crowd – once they go off a party they never elect them again, which means they’ve had five different parties governing them in the last century and political parties regularly die over there:

    http://www.theglobeandmail.com/globe-debate/why-we-might-have-seen-the-last-of-the-alberta-pcs/article24283993/

    Alberta was Harper’s base, so the regional shift in May might mean something…

  35. RAF (and everyone else quoting GB VI numbers)

    Quite why Opinium (and all the other pollsters) continue to provide “GB” headline figures when 6% (a bit high, I think) are for a party operating in the very different Scottish political environment, is a testament to the conservatism of pollsters (and/or their clients).

    Conceivably, it is because the pollsters and/or clients are determined to continue the absurd notion that there is, any more, a single “British” voting climate..

    Alternatively, they just aren’t very bright.

    Despite Anthony’s (rather quaint, if absurd) title for this blog, everyone recognises that there is no “UK” system, and has never been since polling was created.

    As Anthony himself notes, “GB national polls are of so little use in measuring support for a party standing only in Scotland”, though he fails to mention the natural corollary- that large numbers voting in Scotland for a Scotland only party makes the GB figures of less use too.

    I gave the (far more useful) English figures earlier on this thread. Given the low VI for PC, E&W figures would probably be even more useful – and some posters do provide such a crossbreak.

  36. @Bill Patrick

    2020 depends entirely on what Labour do.

    If Corbyn leads them into the election, the Conservatives can safely go in with Osborne and win.

    But if Lab breaks the habit of a lifetime and changes leader and the Conservatives go in with Osborne, it’s anyone’s guess what will happen. Osborne polls very badly (New Labour were very successful in framing him back in 2005, and its a decade later and still that original impression has stuck). Boris would have a better chance, but it again depends on who he is facing.

    Will Lab do what they’ve done for the last 70 years and stick with their leader through thick and thin. Or will it be a 1935 moment when the pacifist Lansbury was removed. It’s too hard to predict at the moment because Lab is such a sentimental party.

  37. @Candy and Laszlo

    Many thanks for the information. As far as i can see from Google, the opinion polls say that the Liberals should be the majority party but are unlikely to win outright while the NDP should come third. The conservatives aren’t expected to be the majority party but I have heard that before and wondered if they would pull off a result. For this reason I am heartened to hear of the sturdy Albertans and that Laszlo thinks it is worth betting against Harper. Hopefully, as far as I am concerned, the bet succeeds.

  38. As this is a polling site, how about we put our heads together to try to formulate some improved questions for pollsters to ask? There are some very knowledgable and expert people in various fields who post here, and not only me :-)

    Here’s an initial suggestion off the top of my head:
    1) Who would you vote for if the GE was tomorrow?
    2) Who would you vote for if the GE was tomorrow and you were £1000 p.a. better off than you are now because of the current government policies?
    3) Who would you vote for if the GE was tomorrow and you were £1000 p.a. worse off than you are now because of the current government policies?

    I’m sure people can pick holes in those questions, but it’s just a starter to get you thinking.

  39. @Candy

    “If Corbyn leads them into the election, the Conservatives can safely go in with Osborne and win.”

    From Opinium – who would make the best PM?
    Osborne 26%
    Corbyn 26%.

    Orborne will not win the next GE for the Tories. He has 5 years to fall flat on his face and has made a promising start.

  40. Won’t 2020 be decided by how well we think we’ve been governed and whether we think the other lot could do any better? We can’t know how well we’ll be governed yet; the polling evidence so far would suggest labour’s quest to show it could do better is unconvincing.

  41. @Candy

    Ehem…Osborne.

    And that’s if he gets the gig. Boris is sniffing around.

  42. Opinium poll provides more evidence that the public aren’t warming to the idea of Osborne as the next PM. I think a grudging respect, if I can put it that way, for him has developed as a result of the economic performance in recent years. However, there is just something about him that people don’t warm to – perhaps it’s that he looks smug, perhaps it’s that he’s seen as too much of a politician and not enough of leader.

    I’m looking forward to Canadian results. I don’t know enough about it to say how votes are likely to translate into seats but a big test for pollsters – much change over course of campaign but Liberals have had clear (around 6 pts on average) leads over the past week or so, benefiting from the NDP seemingly collapsing.

  43. @RAF

    Yeah, but that snapshot was taken in Corbyn’s honeymoon… :-)

  44. @Raf

    Osborne may not be electorally the best leadership option for the Tories. But it would be a massive call to say that that he wouldn’t win given who he’ll be up against (whether or not Corbyn is still around, it is difficult to imagine Labour stumbling across somebody that can satisfy both the selectorate and the electorate).

  45. @Old Nat – You must have been gutted by the penalty. I certainly was. When the rain started to come down I thought that Scotland had won it.

    To be fair to the referee he was not obviously biased against Scotland. He had after all just stopped the Australians from scoring a try by calling for a replay of a fould that most would have missed. In addition it was difficult to tell how exactly the ball reached a Scottish player and who had touched it last. The problem seemed to be that he could not call for a replay because there had been no try and no question of foul play. The referee had to make a split second decision and to my eyes got it wrong

    Personally I think there was more of an issue with the alleged deliberate handball where it seemed to me that the Scottish player was trying to make an interception.

    Anyway which was the worse disappointment the referendum or this match?

  46. Jack and Raf

    Thinking back a few years, we had Heath and Wilson. One was a shifty crook and the other a pompous weirdo. Each of them won at different times, thougha admittedly their policies were much closer than are Osborne and Corbyn’s.

  47. Charles

    I was amazed that we got so close to winning! :(the quarter-final) -)

  48. Opinium

    Corbyn approval -19
    Sturgeon approval -8 Scotland Only +43
    Cameron approval +2
    Farron approval -20
    Farage approval -20

    Only 55% of Labour voters prefer Corbyn to Cameron as PM. That’s not good. 93% of Conservative prefer Cameron.

  49. @Pete B

    I think looking at approval ratings is a good judge of VI. Maybe pollsters shouldn’t ask VI but rather calculate VI from an algorithm based on the supplementary questions.

  50. @Jack Sheldon

    It says something about the state of the Canadian economy that wages there are now seen as cheap compared to the US. Of course no-one in Canada is stupid enough to believe that working in McDonalds entitles you to the equivalent of the UK median wage.

1 2 3 4