The monthly ComRes telephone poll for the Independent is out tonight and has topline figures of CON 32% (nc), LAB 33%(-4), LDEM 9%(nc), UKIP 14%(+4). The one point Labour lead is the lowest ComRes have shown in their phone polls since January 2012, and its the lowest level of Labour support they’ve shown since the government’s honeymoon in the summer of 2010. Meanwhile the Sun politics team have tweeted the daily YouGov poll. That too shows the Labour lead down, in this case to two points: CON 35%, LAB 37%, LD 9%, UKIP 13%. That’s the lowest YouGov lead since December.

As ever, unusual results demand particular caution. Sure, it could be the sign of a narrowing of Labour’s lead, but just as likely it could the random variation that affects all polls. There is a temptation to assume that a movement in the polls after an event – in this case Labour’s 50p tax pledge – is a response to that effect. Labour announce a policy, the next few polls show their lead collapsing – cause and effect. I would urge restraint. At first glance this looks like an obvious and appealing narrative, but it’s a human weakness to look for patterns of this type even when they aren’t there.

Firstly, while ComRes and YouGov happened to both be published at 10pm and show a similar pattern, they aren’t the only polls published today. Populus’s Monday poll was also conducted after the 50p pledge, at roughly the same time as ComRes, and they show Labour’s lead still at seven points. Even without that, we know polls jump about from day to day, YouGov have already shown a couple of 3 point leads this month that turned out to just be normal sample variation.

Equally initial polling showed that the 50p pledge was popular. Now, the reality is rather more complicated than that – a popular policy may play to a party’s wider weaknesses, could risk making Labour look anti-business, or the consequential criticisms from business leaders could have damaged their support. Nevertheless, I’d be surprised if the announcement of a broadly popular policy had backfired that badly.

We’ll have more polls in the coming days – not least we’ll know if YouGov’s daily polls are really showing the lead dropping or if today’s is just a blip. Of course, it could be that other polling does echo these findings and we do conclude that the 50p pledge went horribly wrong, it could be these are just part of a more gentle decline in Labour’s lead that has no link to the 50p pledge at all, it could be that tomorrow’s polls show things back to normal and today was merely a couple of freak results. Wait a couple of days before making a fuss about what could just be a co-incidence.

350 Responses to “New YouGov and ComRes polls”

1 4 5 6 7
    “However, I am not so sure that the proposal is such a vote winner this time because of the levels of adverse publicity”

    The Tories have not recovered their pre Osborne budget polling figures since his 50p top rate of tax cut for the rich…so restoring it certainly won’t be a vote loser. If/when the coalition are voted out of office …this will be the one of reasons why ….. caught looking after the wealthy after the wealthy bankers made the rest of us suffer so needlessly.

  2. Back to polling (sorry to intrude on heated arguments on the 50p rate), but there is one thing about the YouGov poll that I did not notice being picked up at all.

    Some people noted some odd subsamples – a very low 2010 Labour sample, only 39 18-24s, of which only 9 are male. – but it should be noticed that the overall sample was by some way smaller than usual.

    The average sample size for a YouGov about 1860. The range over the past three months or so is between 1650 and 2108. The poll released yesterday had a sample of… 1381 (after weighting – raw sample is 1380).

    It also had a slightly higher proportion of don’t knows and won’t votes than recently, not significantly, but with such a small sample, it means that there were about 1040 actual VIs provided, whereas the average is about 1440.

    This adds to concern about some subsamples to suggest that this poll cannot be as trusted in isolation as other YG polls – and as the normal caveat goes we need to see a few more over the next week to see a trend.

    Even so, it is within MOE of C 32.7, L 38.6, LD 9.3, UK 12.6 – the prevailing averages I arrive at for January so far, weighting by sample size.

    I don’t know much about ComRes, so that poll may be fine or there may be interesting weightings / sample issues there too.

  3. @danivon : the comres poll was also very small, especially after removing the not likely vote pool. there is a link to the data set early in this thread.

  4. OK, so ‘disintegration’ was a bit OTT.

    But even so, for the independence question not even to be (apparently) on the list says a lot about how little folk in the South (I don’t mean ‘small’ folk, but how little it impinges on their thinking) care one way or another. All that does is play still further into Nationalist hands – but if southerners don’t care they don’t care. Nuff said.

    On the other hand, Labour supporters must care (as I have said many many times before on this site) as to lose 35+ seats to an independent Scotland would drastically change their chances of forming a government at Westminster.

  5. Hi TOH, again!

    “The point I was making is that even Labour is strongly divided on the issue and it is not as simple as it seems to you.”

    Fair do’s, but I can only say again, I wasn’t even making a point about raising of tax levels for high earners. You are trying to inveigle me into discussing that!

    I was merely commenting on the sheer insensitivity (towards those in genuine need) of those exceedingly rich people who argue that their situation needs sympathy, or else (at which point they dump the sympathy pitch and turn to bullying) they’ll be out of the country just like that. But they won’t, and whatever else they deserve, it isn’t sympathy.

    Now (seeing as you’ve inveigled me onto the point) suppose tax for the rich was the 60% you once paid. A guy that earns (say) £200,000 p.a. takes home £20,000 out of the last £50,000 he earns, which is better than rather a lot of people earn in total. But the £20,000 person IS motivated by that sum. He looks at those who can’t even get a job, or has a zero hours one, or three part-time ones, or just a minimum wage 40-hour week person, and he thinks, “I’m motivated.” So tell me why a guy earning £200,000 won’t say, “Hey, that last £50,000 is another £20,000 for me. I’m laughing!” And, indeed why the 50% guy, who’ll get £25,000 out of that last £50,000, won’t be dancing in the streets.

    Truly it’s easier for a camel to pass through the eye of a needle than for a rich man to think like a human being, seemingly.

  6. @DANVON

    many thanks for your observations. Any chance that we could establish a rule whereby any poll with a sample of less than 2,500 would simply be ignored as too small to be taken seriously?

  7. Tell me, TOH, is my reply to you of a few moments ago visible to you? After all this time, I still don’t know how this site works!

  8. Headline from The Guardian, probably my favourite of the year.

    ‘GDP figures serve up an ace for George Osborne amid calls for ‘new Balls”

  9. Colin Davis

    No, nothing visible since much earlier.

  10. Paul Bristol

    “after the wealthy bankers made the rest of us suffer so needlessly.”

    As did wasteful Government overspending and the idiocy of ordinary folk, mainly in the US who borrowed recklessly.

  11. TOH

    “As somebody who was paying 60% at the time I was certainly complaining about it. Fortunately Mrs T agreed and it was cut to 40% … Tax at that level was demotivating.”

    As I understood it at the time, there was an argument that the high-earners should be allowed to have more take-home pay. I can sympathise with that argument.

    But where it all went wrong was when top-end salaries went stratospheric. Not only were these people taking home a larger proportion of their earnings, but their total earnings were increasing. It would have been fairer for the country as a whole if their increased take home pay had come from one source or the other.

  12. I heard another interesting stat that household unsecured debt grew a lot faster in the Blair years than now, yet I don’t remember the cost of living debate back then. Was borrowing more acceptable then…

  13. Ah, well, it must be being monitored. No point in trying to put it back. There doesn’t seem to be any kind of a trigger word in it that might prompt moderation, no parties are mentioned, and it’s on a topic people have been discussing all afternoon – and the point I was making was that I was never on that topic anyway. Still, suffice I replied, I guess.

  14. @ TOH

    Yep – and we (they? who?) kept voting for governments which encouraged reckless spending and borrowing on both sides of the Atlantic.

    What really makes me angry, as one who wasn’t reckless, is that I’m having to pick up the tab (the alternative being even worse than not picking it up!)

  15. @RICH

    Yes, good headline.
    If I understand correctly, you are slightly to the right; if Osborne has finally begun to find a series of (long-promised) aces up his sleeve will it make any difference whether Balls is there or not?

  16. John B


  17. @danivon – ” …don’t know much about ComRes”

    Fwiw I’ve looked at their tables and they seem to be saying (page 21) that Labour is leading the Tories by 35%-34% in the South East.

    It also looks like they found too many people saying they voted Labour in 2010, and not enough Tories and Lib Dem.

    Southampton academic Will Jennings, in a twitter exchange with Smithson gives the raw VI as Lab 38.9%, CON 29.0% before weighting: “I predict polling methodologies are going to be in spotlight #GE2015.”

  18. RICH

    @” Was borrowing more acceptable then…”

    Yes Rich.

    At that time everyone thought the merry go round would last forever.

  19. Colin Davis

    It happens to us all from time to time. We start off trying not to be partisan and then somebody touches a nerve and ……………….

    I know you don’t agree with my views generally but it would have been nice to know what you had to say.

    On the polling front the next week will be interesting, will Labours 38% decline, increase or stay steady? Will the economic news have any effect on the Tory 32-33%

  20. @Billy Bob

    At what stage in the lead-up to GE2015 will polling methodologies be put in the spotlight, and by whom?

    Before the GE 2015 we have Europe, Scotland and local elections to get through. How influential are polls? Are they going to influence the run-up (which has already started, some would say) to GE 2015.

  21. I think if anything is going to move VI at all in the near future it’s going to be the Wythenshawe by-election. Nobody expects the Tories to do well or Labour to lose, but the margin of victory, the UKIP vote and whether the Lib Dems hold their deposit is going to make a difference to a few political fortunes, not least Nick Clegg’s.

  22. @John B

    Well, all polls I’ve seen are smaller than that. c. 1000 is not a bad standard sample size, as that gives the 3% MOE with a 95% confidence that we all know and love. 2,500 would reduce that to a 2% MOE.

    I would say don’t trust any single poll, but you can glean useful observations from trends and from averages.

    The question for this one is why it was so much smallera sample than normal YouGov polls.

    @Billy Bob

    Ok, so that looks to me like both samples for the ‘shock narrowing lead’ have question maeks on them and weighting may not be enough to balance out in one, but go too far in the other.

    By the weekend we should have a clearer picture for YouGov, but ComRes will be out there for a while longer.

  23. @John B

    I assume Jennings is talking about after the 2015 result… VI for the general election as opposed to a referendum or lower turnout local/EU elections. He mentions the “changed political landscape.”

    Another person in the conversation (James Morris) says “Comres weight closer to 2010 vote than anyone else. Recall follows vote so as party move away from last election share comres pull them back.”

    Opinion poll results are obviously used selectively to drive a particular media narrative, and non-VI questions generally are used to test particular messages.

    I would guess there is a percentage of the population who will tend towards backing the favourite… and any number of non-political media figures who will instinctively try to accomodate themselves to any likely changing of the guard.

  24. @DANIVON

    Thanks for your reply.

    If 1000 was a reasonable sample under ‘normal’ circumstances, in the days of ‘uniform swings’, how does that stack up against the realities of the situation in which we find ourselves now?

    1000 for the UK (or GB, as NI doesn’t really enter into normality) that still give only c.85 respondents in Scotland and c.50 for Wales. Is that enough, given the complexities of the situation?


    “As did wasteful Government overspending and the idiocy of ordinary folk, mainly in the US who borrowed recklessly.”


    Well yes, many got suckered into the sub-prime mortgage thing in the States. But… there was a fair bit of misselling going on. And your career may have involved finance, so it’s easier to be up to speed on these things, but for the rest, it can be a minefield, and these days they make it ever more so. Many things are more complicated now than in the baby-boomer heyday.

    And for people having to hold down two jobs, even harder. Add in all the info. you need to take into account these days on health, education, all the technology we depend on, and much more besides, there aren’t enough hours in the day. So we depend on the specialists.

    If that’s not enough, bear in mind that when it comes to the sub-prime thing, even the financial specialists, the banks themselves, got taken in and bought these dodgy mortgages, the toxic debt that brought the banks down. So it’s no surprise if the lay person struggles…

    And again, this highlights the double-edged sword of incentive. If someone like Dyson is getting rewarded, that’s one thing. I doubt there are many on this board who would complain over-much that he’s building a new facility and planning on hiring three thousand engineers if I recall correctly.

    But when it comes to the purveyors of toxic debt, or the bankers who bought it by the bucket-load, and who made bundles regardless, maybe that’s incentive we can do without?

    On the bright side, you got your wish to send your offspring to private school, and part of the reason those institutions still exist, is because of the taxation that allows them to have charitable status, and funds state education, thus keeping people from feeling things are even more stacked against them. So it’s tolerated…

    These days of course, fees have rocketed in the better schools. People paying less tax, so schools can charge more. The wonders of demand pricing again…

  26. Colin D

    The AutoMod will not tolerate any suggestions of mendacity. To the extent that, if your post includes, for example, the word “comp-l-ying” it wi be cast into the eternal darkness.

  27. TOH

    @”contruction is notoriously volatile and the figures may well be revised upwards in the future.”

    Absolutely-the ONS record shows how volatile it is month to month, and notoriously difficult to track; construction GDP revisions can be very large.

    And these first estimates are based on two out of three months numbers-December in this case being guestimated.

    I really don’t know why ONS report data on a month on month basis. US GDP stats are reported on a yoy basis.

    On a quarterly basis Construction GDP was up 4.5% yoy in Q4 2013.

    And reports like this don’t persuade me that we are seeing a contraction of the sector :-

  28. Interesting post Danivon – thank you.

    A question that I have long wanted to ask is why are the sample sizes so small? Has it been established that there is no advantage … for example in reducing the MoE?

    I don’t know what logistical/financial problems there might be in increasing the sample but I can’t help but think that an awful lot of the discussion hinges on the drift inherent in ‘n’ being too small.

    There is no doubt that the reason I’ve always valued Yougov is that instead of sporadic one-offs, their polls identify trends. A point that Adam Boulton (and his other media chums) seem to find inconveniently get in the way of their high octane stories.

  29. @MrNameless

    I see that some people think that UKIP will win in Wythenshawe & Sale East.

    Anything like a result along those lines (UKIP at above 20%) will create shock waves.

    Goggins was a good man. It would be sad to see his legacy wrecked.

  30. They’ve got no hope, frankly. They’ll probably take second, but distantly. They did pretty badly in a council election last week and their candidate just went broke.

  31. @Danivon

    I see that you’ve already answered my questions.

  32. @ Mr. Nameless,

    Do you really think the Lib Dems losing a deposit in a northern Labour seat would matter? I would think they’ve grown used to it by now.

    The only eventful outcome I can imagine for this by-election is a Ukip win, and as you say the chances of that seem remote.

  33. Yep, I’ve got a post that triggered auto-mod too, if AW would be so kind as to rescue it.

    Incidentally Anthony, is there a chance that we might be able to post images of polling-related graphs and stuff? Obviously they’d need to go into auto-mod for vetting, but it’d be quite useful for polling-related discussion…

  34. @ Rich
    “household unsecured debt grew a lot faster in the Blair years than now, yet I don’t remember the cost of living debate back then.”

    There was a huge increase in consumer credit which was criticised at the time by some commentators — not Tory ones — as an excess of free-market credit policies which was storing up trouble for the future. .

    There was less debate about the cost of living because real earnings rose rather than fell & because policies like the Minimum Wage, higher child benefit & Tax Credits boosted the incomes of the lowest income earners under Blair & indeed Brown. Voters in the 2nd quartile did less well which is presumably why the swing to the Tories in 2010 was highest among C2DE voters.

  35. Colin

    Thanks for construction reference. Useful.

  36. Syzygy –

    It’s a trade off between cost and usefulness.

    Larger sample sizes give increasingly smaller benefits in terms of precision – doubling the sample size does not make a poll twice as precise: a 1000 sample has a MOE of 3%, but a 2000 sample has a MOE of 2%, not 1.5%.

    Newspapers, frankly, have very little money to spend on polling and there is very little money to be made in political polling anyway. It’s not an area where there is lots of money sloshing about to spend on bigger samples.

  37. @ Carfrew
    If you buy the new Poll Doll, de luxe model, it enunciates all the words which trigger auto-mod. It does sound like an infant Dalek but what do you expect from a Pound Shop.

  38. Mr Nameless
    You quoted:
    IpsosMORI Issues Index (January):
    What do you see as the most important issues facing Britain today?
    Race Relations/Immigration: 41% (+4)
    Economy: 41% (+2)
    Unemployment: 32% (+3)
    NHS: 27% (+5)
    Poverty/Inequality: 16% (+2)
    Crime/Law and Order: 16% (+1)
    Pensions/Benefits: 16% (+6)
    Inflation/Prices: 15% (-2)
    Education/Schools: 15% (+3)
    Housing: 11% (-2)

    Thanks and interesting.

    I note that there is no prompting, people just name what they think and Mori sticks it in an appropriate row.

    Looking at the tables which are here (pdf)

    I did wonder whether a priority would be evident but no such, so I suppose people just went on naming issues until they were exhausted. It would have been interesting to have had an insight into that sequence.

    I wondered why EU was not up there and found it down at 65 out of 970 weighted answers (7%). In the age groups, 45 of the 65 were 55 plussers.

    A gold mine for canvassers. Pity about the prioritising, but the non-prompting is to be applauded.

  39. TOH

    That Building Design link didn’t open up the article I think?

    This is a copy :-

    “The value of construction projects in the UK rose by 67% in the last year, according to the latest figures.
    New project contracts awarded in December were worth a total of £6 billion, compared to £3.6 billion in the same month in 2012.
    The infrastructure sector performed particularly strongly, with the total value of projects awarded in December reaching £2.2 billion, 10.6% up on the previous month and more than 100% year-on-year growth.
    The figures, from Barbour ABI, also reveal sustained growth in the residential construction sector. The total contract value for December in this sector grew by 88.2 per cent year-on-year.
    Michael Dall, lead economist at Barbour ABI, said: “Residential construction remained strong throughout the latter half of 2013 and has truly spurred on the growth of the industry.
    “In the most part, the increase in residential activity can be attributed to the government’s Help to Buy scheme, but there are concerns that demand created by such schemes will surpass the rate of house building. This is a topic that will be the source of major debate over the next 12 months in the lead-up to the 2015 elections.”

    “The performance of infrastructure as a sub-sector is also incredibly encouraging. The figures are a reflection of an exciting year of renewed optimism for the construction industry.

    “As a major contributor to figures for the industry as a whole, the sustained period of strong project values in this sector is good news and a key indicator for future growth.”

  40. Robbiealive,

    “There was less debate about the cost of living because real earnings rose rather than fell”

    Which makes no sense, because if wages rise by 1000% and prices rise by 1%, the cost of living has still increased.

  41. @Spearmint

    “Do you really think the Lib Dems losing a deposit in a northern Labour seat would matter? I would think they’ve grown used to it by now.”

    It’s not exactly unprecedented, no, and it’ll make smaller waves than it might once have done. But it’s another straw on Clegg’s back and not well timed after Rennard and Hancock incidents.

  42. @ AW
    “[Polling’s] not an area where there is lots of money sloshing about to spend on bigger samples.”

    Well it exists in the real world then.
    There are also v. few paying ads on this site, tho these must cluster at elections etc when the no. of visitors increases. rather than in periods of anorak aridity.

  43. Thanks for unmodding my 6:51pm post to TOH, Anthony.

    In the pub reading the Independent, the headline is about the Libdems “Declaring war on the rich”. 2015 manifesto plans to raising 2Bn from a mansion tax, and a.pledge to reduce reduce tax-free limits for pension pots by 20%. (Of course, these are LibDem pledges, but still…)

    Given the 50% tax positioning, it seems like “all in it together” is shaping up to return to the electoral battleground, alongside cost of living, welfare, immigration, economy etc.

    Be interesting for polling to compare the salience of AIIT (All in it Together) with the other hot potatoes…

  44. @ Bill Patrick re Rich.

    I used the phrase @ Rich used, by which he presumably meant real living standards, as I am sure you are perfectly aware.

  45. @Robbiealive

    Lol, at times one suspects the list of words which don’t trigger automod might be briefer…

  46. Anthony,

    Just curious. I can understand why adding more people into a telephone poll would add significantly to the cost, but is it the same issue with Internet-based polling? I assume everything is handled by computers rather than someone having to go through the responses individually, so where does the extra cost come in?

  47. Alex F,

    Polls have to be run for longer to get the requisite larger sample – this means running more polls concurrently, which put strain on servers as well as annoying clients who want their polls in a timely manner.

  48. @ Carfrew

    “Lol, at times one suspects the list of words which don’t trigger automod might be briefer…”

    Dunno I’m pretty much a stranger to auto-mod.

    As I used to say to my daughter “List the things I CAN tell your mother; it must be shorter than the list I can’t!”

  49. @TOH:

    I don’t think it’s true to say the 50p rate has been strongly opposed within the Labour Party (if at all). I certainly wouldn’t include Digby-Jones. Although briefly a junior minister under Brown he declined to join the Labour Party and has since flirted with both the Tories and UKIP (also described as ‘the walking personification of the spirit of big business at its corporatist worst’ by the Telegraph’s Iain Martin). Paul Myners, the other former Labour appointee, is also not a party member. Those who’ve supported the proposal include Alastair Darling and Andrew Adonis.

  50. @RogerH @TOH

    Re: 50% tax

    I suspect that if Labour Party members were directly asked if they support the policy, the yes vote would around 90% at least.

    That would be a comprehensive endorsement in anyone’s language.

1 4 5 6 7