We have three new polls so far today. TNS have put out a new GB poll, which has topline figures of CON 33%(nc), LAB 32%(nc), LDEM 8%(+1), UKIP 16%(-1), GRN 5%(+1) – clearly no significant change since their previous poll (tabs are here).

ComRes have a new poll of the 40 Labour held constituencies in Scotland (that is excluding Falkirk, where Eric Joyce sat out his term as an independent). In 2010 the share of vote in these seats was CON 14%, LAB 51%, SNP 19%, LDEM 14%. The new ComRes poll found support standing at CON 13%(-1), LAB 37%(-14), SNP 43%(+24), LDEM 2%(-12). The seven point SNP lead represents a swing of 19 points from Lab to SNP, the equivalent of a sixteen point SNP lead in a national Scottish poll (tabs are here).

Finally YouGov have a new London poll for the Evening Standard, which has topline figures of CON 34%(+2), LAB 45%(+1), LDEM 8%(+1), UKIP 8%(-2), GRN 4%(-1) – changes are from YouGov’s previous London poll a month ago. The eleven point Labour lead represents a swing of 4.5 points from Con to Lab since the general election, the equivalent of a two point Labour lead in a national GB poll (tabs are here).

444 Responses to “ComRes in Scotland, YouGov in London and latest TNS poll”

1 2 3 4 9
  1. @Alan

    For some reason (time lag?) Con seem significantly behind Lab on the current polling. I think their factor to stabilise the VI is doing something odd. A few weeks ago we saw a crossover which then went away as if it never existed.

    Yes – for some time I have been perplexed about the calculations that lie behind this graph. Some weeks ago @Spearmint and I exchanged comments about events that showed up and then disappeared. The oddest thing is that the graphs seem to be subject to retrospective revision. Today’s graph shows a time a couple of weeks back when the Tory and Labour VIs touched each other. A couple of days ago (that is, well past the convergence date) the Labour VI was shown as being ahead at this point. So, even though it was ten days past at that point it has since been revised from the initial plot.

    This could happen if they are careful about recording fieldwork dates. A poll may turn up based on fieldwork that is quite old and so may alter figures previously posted.

    Like you, I can’t understand why they still have Labour almost 1% above the Tories. I seem to remember reading something about assigning different weights to different pollsters. In this case, it may be that their trusted pollsters are still showing a Labour lead. All a bit impenetrable…

  2. @ProfHoward

    Sounds like PK’s third prediction in a week.

  3. @OldNat

    Not sure if you still follow such matters but we’ve missed our usual tri-annual (every 3rd year) SIMD update for 2015 because the revised datazones weren’t agreed until November 2014, after the stats collection date of September 2014.

    Instead it will be based on September 2015 data and published in Spring next year as the 2016 SIMD.

    With the new (in my area much more accurate geographically due to hours of work responding to the consultation!) datazones I suspect we will see some very interesting changes.

  4. @Omnishambles

    . If you’re checking the numbers regularly, how often are those VI numbers changed? Because I could have sworn it was 13.3% for the LDs about a week ago.

    New polling data is added at least daily (at a time specified at the top of their home page). As a result changes ripple through all of their tables and projections. The LD figure was actually 13.6% a week or so ago. (I posted a similar comment to the one eralier tihs evening)

  5. @RAF 10:03

    Per OLDNAT 8.58 post based on analysis of ComRes data

    In the middle group of ca. 12 seats there is a dead heat between SNP and LiS.

    This is the rump that LiS seem to have the best chance of protecting. If SNP are better at getting their vote out than LiS then even holding on to these is going to be a tall order – especially if Lab resources get diverted south.

  6. Unicorn
    They’ve stopped putting the time. Just the date now. So it’s a bit unclear to me exactly which polls are included.

  7. @pHoward – One of the key lessons of that ST YG poll in my mind is to illustrate that no one – not even the admirable Peter Kelner – is immune from a bit of gun jumping when noteworthy polls appear.

    He moved from Lab ~263 to 297, now back to Tory minority.

    I think prediction is a mugs game this time around, but it will be fascinating to see who gets close to the result.

  8. @RAF

    Sounds like PK’s third prediction in a week.

    Yes – he seems to have reverted to a rather long-standing earlier projection after a very uncharacteristic wobble on Sunday (apparently based on the results of a single poll).

  9. @PeteB

    SIMD (and similar measures used down south) are quite poor at picking up deprivation in rural areas.

    SIMD is a measure of pockets of concentrated deprivation and so doesn’t pick up areas of marginal in work poverty in rural areas where poorer people live alongside other more affluent neighbours.

    For funding / targeting of programmes we usually apply an additional rurality measure to compensate for the poor performance of SIMD in rural areas.

  10. ExileinYorks

    There are worse places to be an exile in!

    The turnout question was raised on the previous thread too. It’s one of the great unknowables.

    We can speculate as to whether supporters of the SNP are more motivated to turn out, despite more of them belonging to demographics traditionally less likely to vote.

    But we won’t know till the results come in.

  11. @Alec


  12. How does ComRes justify polling in Scotland and just listing Labour, Conservative and LibDem directly and lumping all the rest, SNP, UKIP, Greens, into the OTHER category? This is especially misleading as most current polling has SNP, UKIP, Greens at over 55 percent of the Scottish electorate.

    Even worse, this was not a UK or GB wide poll, but a specific poll aimed at determining the relative status of 40 SCOTTISH constituencies only.

    Even in 2010, the SNP was the second largest party in popular vote in those constituencies.

    This appears to be another example of London-centric bias in polling.

    I wonder how people in SouthEast England would feel if the party choices were Conservative, LibDem, UKIP and Others or in the North of England if they listed Labour, LibDem, UKIP and others?

    I am sure many would be upset.

    Often on this site, people wonder why Scottish-based posters are anti-London or anti-Westminster. This type of action is more common than not and explains a lot.

  13. Unicorn

    That was the tentative conclusion I came to. I suspect there might be some maximum likelihood smoothing algorithm that retroactively changes the paths.

    The fact that it doesn’t seem to be converging is very odd if it doesn’t reflect the polls at the end it’ll be a very weird prediction.

  14. Northumbrianscot

    I’m delighted to say that I have had to take little detailed interest in SIMD and similar measures since I retired. :-)

  15. Additionally, how can ComRes be sure they have a representative sample of Labour v SNP voters?

    The 2010 other group which is used for weighting includes SNP, Green, UKIP and is down weighted considerably. How are the voters assigned back to the various parties in that group?

  16. @Unicorn – to be fair to PK, I gather he did suggest the underly!ng leadership questions were part of a shifting picture, and these have been confirmed in other polls showing EM’s ratings climb.

    I’m sure AW suggested that this was pretty much all from Lab supporters anyway, but it’s still potentially important. Turnout is going to be particularly important this time around, and if Ed can solidify Lab voters and convert more to actually vote, this possibly helps as much as winning converts.

    [Wasn’t me – it’s something I’ve been meaning to look at, but haven’t had chance – AW]

  17. OLDNAT

    I have to agree we won’t find out until the results come in. It’s such uncharted territory that neither past patterns, nor current polling data are going to shed much light on it.

    For me it is why the way Scotland votes is going to pivotal to this election.

  18. OldNat/James

    We do actually have some indication of Green areas of strength from Ashcroft polling, though irritatingly they’re not summarised in his report, but lumped under Other. The one that immediately comes to mind id Glasgow North where the Green CVI vote was 11%. Of all those Greater Glasgow seats, North was the one with a high Lib Dem vote in 2010, which in this poll collapsed from 31% to 4% and, as you would expect, the Greens benefited from that, though as elsewhere, a lot of their support came from new voters.

  19. By the way, Newsnight last night had a seat count of pro-Tory MPs as 319 and pro-labour as 319, with the projections done by the Professor at UEA.

    They say they plan do do this each night.

  20. @OldNat

    I suspected as much!

    I’ll maybe come on here next Spring and let you know how the 2016 SIMD is looking ;-)

    I have to say I think SIMD is overused by funding bodies because of its flaws in picking up sustained but I concentrated poverty but it is an interesting (if you like that sort of thing) snapshot of an area.

    It doesn’t really tell you anything a half decent Officer shouldn’t know anyway from having been out on the ground visiting neighbourhoods and schemes.

  21. YouGov/Sun poll tonight – Labour lead by one: CON 35%, LAB 36%, LD 7%, UKIP 12%, GRN 5%


    YouGov/Sun poll tonight – Labour lead by one: CON 35%, LAB 36%, LD 7%, UKIP 12%, GRN 5%

  23. With differences

    atest YouGov poll (30 – 31 Mar):
    LAB – 36% (+1)
    CON – 35% (-)
    UKIP – 12% (-)
    LDEM – 7% (-1)
    GRN – 5% (-)

    Comment: as we were. Confirms neck and neck. But puts top two on a high joint score.

  24. Interested in the Lib Dem tally in the Comres Scottish Labour held seats. This shows them down 11% at 2% compared to 2010, forming the equivalent of perhaps 40% of the SNP increase.

    Has their been any more detailed analysis of former LD voters in these seats, as I would have thought that there would be an opportunity for Lab to target these to close the gap on the SNP?

  25. I thought Con would lean in tonights YouGov. Combined vote for Lab/Con up to 71%. There is definitely a squeeze going on here.

  26. @ Brian Nicholson 10:25

    I don’t know whether to laugh or cry (at ComRes) ….

    Seriously though, It’s just sloppy and doesn’t inspire confidence in their results.

  27. Before bidding good night; 36 for Labour is high as is the 7 for Lib Dems.

    The race is real.

  28. Hello 71%

  29. @Pete B, Alan etc

    Perhaps the Labour advantage is not all that strange. In the relevant section of poll-use methodology they state:

    We then take into account that pollster’s general tendency to have certain parties at higher or lower vote share relative to other parties (that is, house effects, see the current estimates of these here). Finally, to estimate how much variation we can expect from poll to poll, we also take into account the typical number of respondents giving a party preference in that pollster’s polls, which varies substantially between pollsters (telephone polls typically have samples in the 450-750 range, online polls in the 750-1500 range). We use the number of weighted respondents reported by each pollster in their data tables, before any allocation of respondents who said they did not know how they would vote.

    This mentions at least two corrections that make things different from straightforward rolling averages. First, they adjust the polling results for known house effects. Secondly they take account of sample size (which not many of the rest of us do).

    Over the last few days, many contributors decided that the YouGov +4 for Labour was cancelled by the ComRes -4. But the latter had a smaller N, so won’t have carried so much weight.

    Taken together these two adjustments could well explain why they still judge Labour to be slightly ahead.

  30. I reckon the rolling average is at 34/34 now.

  31. @Alec

    Anecdotally the 2010 Lib Dem voters in traditional Labour seats tended to be Old Labour types who fell out with Labour over Iraq / Blair but weren’t yet convinced by Independence.

    The evidence from 2011 is that a lot of these voters crossed the floor straight to SNP so I’m not convinced they are a fruitful source of votes for LiS.

    Ex Lib Dems in seats like Edinburgh South with large numbers of English Born and/or middle class voters may be more fruitful territory.

    Ian Murray could plausibly go from being LiS’s most marginal MP in 2010 to having LiS’s largest majority.

  32. So, another biggish (2K?) polls showing a Labour advantage. Perhaps that balances out a couple of smaller (1k) polls going the other way (e.g., Ashcroft and TNS).

    Perhaps we should all start being a bit more sophisticated when doing our averaging…

  33. YG five poll rolling averages:

    Con: 34.4% (like yesterday highest of the year)

    Lab: 35.2% (highest of the year)

    LD: 7.6% (second highest of the year)

    UKIP: 12.4% (lowest of the year three days in a row)

    Green: 5.4% (lowest of the year)

    I could type the same thing every night with the current YG polls…..

  34. @ Brian Nicolson

    Additionally, how can ComRes be sure they have a representative sample of Labour v SNP voters?

    They can’t be sure. But nor can any of the other polling firms.

  35. Northumbrianscot

    “I’ll maybe come on here next Spring and let you know how the 2016 SIMD is looking ;-)”

    I’ll take a benign interest. :-)

    I was one of the Officers (then Consultant) trying to explain the numbers to Councillors, when SIMD slowly penetrated their skulls. The difficulty was not exploding into “How the **** could you not have known what was happening in your ward?”

  36. Everything is still within the margin of error of Con 33, Lab 33, UKIP 13, Lib Dem 8, Green 6. Con and Lab may be inching up and UKIP down but I think any movement is minimal so far. It may be different by next week.

  37. @ Northumbrian Scott

    Ian Murray could plausibly go from being LiS’s most marginal MP in 2010 to having LiS’s largest majority.

    I’m not sure about the largest Labour majority in Scotland but I do agree that Ian Murray will hold, probably with an increased majority.

  38. In the G the LDs are proposing US style First Amendment rights of free speech.

  39. interesting that labour still hold a lead on YG…I will be interested to see if YG have crossover before the election….

  40. @BazinWales

    These trends have been running solidly for about a fortnight, with good evidence to support them.

    They could reverse before the GE off course, but they are more than just MOE. They look more solid than that.

  41. RAF

    As long as the freedoms aren’t Indiana style!

  42. @OldNat

    Agreed that the value of SIMD is it provides “proof” of what should be obvious to a decent Officer / Councillor.

  43. ‘of course’ not ‘off course’.

  44. @raf

    That’s interesting.

    Important difference being that it’s a lot easier to change an Act of Parliament than change amendments to the US constitution… we’ll never have something like the 1st amendment unless we codify our constitution

  45. @OldNat


  46. @Unicorn
    “Perhaps the Labour advantage is not all that strange. In the relevant section of poll-use methodology they state:….”

    Is this about Election forecast? If so, though your contribution was very detailed and expert, I thought that EF had Conservative ahead?

  47. @Omnishambles

    Indeed. In the US, you need 2/3 of voters in 3/4 of states to approve the change (or vice versa – I forget!).

  48. I am beginning to wonder if Labour might be regaining a little ground in Scotland. If not, they must be doing well in England in terms of swing.

  49. “I still reckon we are heading for Con having the most seats but well short of a majority and unable to form a coalition.”

    this increasingly looks like a blind statement of faith, made slightly more bizarre by your belief that labour will do better in scolland than current polls indicate.

1 2 3 4 9