There are two new Scottish polls in Sunday’s papers, an ICM poll for the Scotland on Sunday and a Panelbase poll commissioned by the Yes campaign. These are only the second and third polls that we’ve seen conducted wholly after the debate between Alex Salmond and Alistair Darling, the first one from Survation suggested a significant move towards NO, but these two paint a different picture.

Taking the Panelbase/Yes Scotland poll first, topline figures are YES 42%(+1), NO 46%(-2), Don’t know 12%(+1). Excluding don’t knows this works out at YES 48%(+2), NO 52%(-2). Changes are from the Panelbase/Sunday Times poll in July. Compared to last month it’s a slight move towards YES, but is not particularly significant by Panelbase’s standards: their penultimate poll also had YES on 48%, and they’ve have YES on between 46-48% since March.

The ICM poll in the Scotland on Sunday has topline figures of YES 38%(+4), NO 47%(+2), Don’t know 14%(-7). Excluding don’t knows this works out at YES 45%(+2), NO 55%(-2). It’s a bigger NO lead than Panelbase are showing, but the same modest movement towards YES. As with Panelbase, it’s not a massive change from the longer term trend – so far this year ICM have had YES between 43% and 48%, and YES 45% is right in line with the average of all ICM’s Scottish polls this year.

So, two polls, both show a modest movement towards YES since last month, but neither a significant shift from the longer term trend. What it does mean though is that the movement towards NO in Survations’s post-debate poll has not been echoed in other companies’ polls.

(Incidentally, I’m on leave this week, having a break before the long slog to the general election, so expect light blogging for the next few days)

198 Responses to “New Panelbase and ICM Scottish polls”

1 2 3 4

    Interesting analysis on the tendency in referendums for the support for change to fall away between one month out and the actual result.

    However, in the analysis, there is a little to keep everyone happy.

  2. Floating Voter

    “inflation expected to be just below the 2% target by the end of 2014, average real pay is expected to fall for the rest of the year, with rises not expected until 2015.”

    I cannot understand why anybody should be in the least surprised about that. It’s the price to be paid by the nation for living beyond it’s means for years and years under all governments. Until the size of the state is cut back significantly it is likely to continue. It may well effect the outcome of the next election although I doubt it but a change of government is unlikely to make any difference anyway.

  3. @Alec

    “Consumer confidence at new high” The Telegraph

    “Lloyd s Bank showed consumer confidence reached 147 points in July.”


    “It said confidence had increased in step with improved spending power, ”


    “Lloyd s’ wider gauge of confidence in the current situation meanwhile continued a seven month unbroken rise to hit 181 points. Survey respondents were also more positive about their job prospects than at the same time last year..”

    However according to Springboard

    ‘total retail footfall recording a 0.6 slump last month. ‘

    i read similar from the British Retail Consortium for July

    So there we are – consumers more confident but spending less.

    We can find out the real situation when the ONS reveals the inflation figures on Tuesday and retail sales on Thursday

  4. @The Other Howard

    “a change of government is unlikely to make any difference anyway.”

    I think that your last sentence may well be correct and very perceptive.

  5. Dao Dao

    Thoughtful contribution but three polls from three different organistions in two days showing movement towards YES is worthy of note and probably real change.

    One unusual problem for NO is this. Any miniscule movement towards NO is taken by the old written press as a massive collapse in YES support. This is the press that most posters here still read and are influenced by.

    Significant movement towards YES as we are now seeing goes unreported or ignored breeding complacency and extraordinary arrogance in the NO camp.

    All of this old press propaganda (such as the idea that Darling had a great debate for example) has little effect on the people who believe what they are hearing from friends and neighbours not the Telegraph or the Mail.

    The momentum is now with YES, and while no doubt there will be further bumps, along the way it may well carry them to success over the next month.

  6. TOH

    @”t a change of government is unlikely to make any difference anyway.”

    Its the thought which occurs to me every time EM uses the average pay statistic. But how many people will ask-well how will you change it ?

  7. Carney at it again. He writes in the ST that BoE may increase rates before average pay rises.

    Markets “surprised” & sterling rallies.

    His ever changing forward guidance is becoming a bit of a joke.

  8. Populus:
    Lab 37 (+2)
    Con 32 (=)
    LD 9 (=)
    UKIP 14 (=)
    Oth 8 (-1)

  9. FV

    @”So there we are – consumers more confident but spending less.”

    Footfall isn’t the same thing as spending.

    “”However, strong footfall doesn’t always equate to good sales and vice versa. Retailers have seen good sales growth in their online clothing and footwear offer and many are reporting positive results led by their online summer sales. The popularity of click-and-collect has helped support footfall in retail parks offering further evidence of the changing shape of how we shop and the increasing complexity of the relationship between digital and physical. ”

    Helen Dickinson

  10. @Colin

    “Footfall isn’t the same thing as spending.”

    Yes I made a false assumption there, and I was forgetting about online sales

    We will still have to wait for retail sales figures on Thursday from the ONS

  11. To return to the question that I asked earlier about Panelbase’s problems of a year ago, no-one has responded with any evidence that their panel has reopened. It was frozen a year ago, when evidence emerged on social media of nats encouraging each other to join the panel in order to influence referendum polling results.

    So what are the implications of this for current polling?

    Firstly, Panelbase stated at the time that ” new joiners have had no significant effect on our results”. But I don’t find that as reassuring as it is perhaps intended to sound. There would be have to be quite a marked variation in results between those including and excluding new joiners before the difference satisfied a test of statistical significance at a 95% level. It’s little different to our inability here to dismiss the potential for m.o.e. when commenting on variations in successive polls of a couple of percentage points or so. So in practice, that statement gives quite a bit of potential for wriggle room.

    Commercial interests may also come into play. Look at the alternative. Say Panelbase accept that there is a problem with their panel which they can’t easily put right. That’s akin to throwing away a year’s lucrative referendum polling commissioned by the “Yes” camp as their main client. Does that big financial penalty offset the risk that their polling might be biased by say a couple of percentage points in the direction that their main client would want anyway? It seems to me that in this case there was quite a big commercial incentive to try and brush problems under a carpet.

    And in reaching a judgement on whether or not Panelbase can be trusted, consider also their response when they were widely pilloried by asking a prior question that would very clearly have biased their main referendum polling i.e. “We accept that it is not inconceivable that this may have had some limited effect – although we cannot really know for sure”. In other words, doing their level best to play down the impact of a different problem which they couldn’t really deny. That record of playing down an undeniable problem is a reason to be more sceptical when they issue a statement denying a “significant” impact of entryism on their panel.

    Leaving aside the issue of trust, there is one other consideration too. As Panelbase’s panel was frozen a year ago, some on it will have since ceased to take part in surveys and response rates will be down. Without the normal churn of replacement new entrants, it seems reasonable to expect that those who have not got tired of responding to surveys include those most keen to have their views on the referendum known. And such differential response rates are just the sort of things that pollsters normally want to avoid.

  12. Some weird stuff yesterday mainly but not just from our very own contingent of cyber-nats. Darling will be in for another kicking from AS next Monday, yes I know, but even if AS loses he wins ! According to the usual s. I can’t wait to see the debate,
    The poll looks about right to me, although CL might differ.

  13. The populus poll that is.

  14. FV

    Retail is changing rapidly.

    ONS’ numbers for year to June 2014 are illustrative :-

    YOY% +/- Volume/Value

    Retail Sales +3.6/+3.5
    Fuel +0.4 / -1.9
    Stores-Food +1.1 / +1.9
    Stores-Non Food +4.6 / + 4.5
    Non-Store +17.1 / +16.5 ( !)

  15. @ Colin

    Totally agree with your views on “forward guidance”. It has become a joke. That’s not to say that the decisions being made are a joke- it seems sensible to not raise interest rates until any recovery is established but he chose the wrong indicator (unemployment) as the guidance. In fact he should have been looking at pay rises which would have reflected a better guide to the state of the employment market.

    I’m not an economist but I am beginning to wonder what the downside to permanent low interest rates is. On paper it should aid the economy by encouraging investment and should also help domestically with the biggest cost in anyone’s life of buying a house. I know it affects pensions and savings but technically low interest rates should be beneficial for companies performance on which much of your pension pot is based.

    Unfortunately, in my humble opinion, the financial markets are so distorted with the emphasis on speculation rather than investment that it probably has little effect. Certainly with house prices (caused by a land/housing shortage) low interest rates make little difference because you just end up paying more for the house in the first place as anything with a shortage is priced to what people can afford not what it takes to build.

  16. When I wrote that Populus would probably give a ‘ten point lead by Monday’ I was actually joking. Nevertheless, it ‘only’ went up to 5, so it proves what a hopeless forecaster I am, which is why I don’t forecast.

  17. Colin,
    My job involves buying from Eire and selling in mainland Europe and last week I was waxing lyrical with colleagues on Mark Carney’s inconsistency leading to greater fluctuation in the Eu/£ than necessary affecting our decision making.

    My colleague (retiring soon) dryly suggested that the Governor of the BoE probably did not need an lesson in Economics from me to much amusement all round.

    I rather weakly mumbled that it is PR that he needs lessons in – he may well be a great Economist, I don’t know – and that comments by the Governor of the Bank of Canada meant little but in this job he needs to be circumspect.
    Whilst Eddie George was a tad political imo before the last GE his steady Eddie approach did not confuse the market at least.

  18. Niall Patterson of Sky News being very positive of the Yes campaign. Patterson said AS did well in the debate with undecideds, that the trend across all polls is towards Yes, and that the campaign really starts now.


    I don’t blame him for searching carefully for the right moment to tighten monetary policy. I am encouraged that he is thinking about the implications for all sectors of the economy.
    But I think he made a mistake trying to find precise pegs & measures. And to keep moving on to the next criterion just confuses everyone.
    He is also far to ready to voice policy outside the MPC forum & BoE official announcements. Andrew Tyrrie’s Select Committee has already taken him to task on this.

    SHEVII-its an interesting point-what’s wrong with ultra low rates forever> I’m not an economist , but I imagine the answer lies in the various balances in the economy-savers/borrowers-importers/exporters-etc etc. And of course when BoE stop re-investing QE gilt redemption receipts-or even start selling pre maturity, market rates will start to rise anyway.

  20. JIM JAM

    I thought saw you post a comment that you were revising your view of UKIp to Con returnees.

    Mind sharing those thoughts ?

  21. @raf

    If that summary is correct, what a facile analysis. That’s the stuff of journalism though isn’t it? That’s why I come here, to get some facts…

  22. @Colin – “His ever changing forward guidance is becoming a bit of a joke.”

    I do recall I said as much a long time ago (once he first backtracked on the unemployment rate) and you gave me something of a ticking off, supporting MC.

    I remain of the opinion I have held of Carney right from the start – I’m unconvinced that his record gives us sufficient confidence in his abilities, the entire concept of forward guidance is a mistake, and there should have been a micro adjustment of rates upward months ago to remind consumers that this won’t last without frightening them.

  23. Colin – my view has been that UKIP would end up with 6% ish with the Cons receiving a net boost of around 3% over Labour – using YGs 12%.

    CB and I have disagreed with him reckoning on 10-12%.

    I am wondering whether the UKIP vote may be more resilient than I thought and than 8-10% could be possible with a smaller net gain for the cons in overall VI.

    I also say in the same post, though, that these UKIP votes will be mainly in safe Labour and Tory seats.

    The bombardment of vote UKIP let in Lab/LD in (Tory in a fewer cases) messages in marginals will be difficult for anti-EU kippers for example (2010 Tories) to ignore.

    In essence the UKIP may do better than I have thought likely up to now but the significance will be limited in seat terms.

    Finally – not revised yet but like Fisher et al if swing-back dragging its’ heels I have to be willing to reconsider.

  24. Some interesting numbers in the housing market.
    Nationwide reports mortgage lending down & (deposit taking up)
    It cites BoE’s Mortgage Market Review – new rules under which lenders have to ask borrowers more detailed questions to check whether they can afford their loan.

    Rightmove reports average asking price of new sellers in England and Wales was 2.9% lower than in July, with London prices down 5.9%.
    Spokesman said :-” expected to see more price falls next month before the usual autumn flurry of activity. “It will, however, come as some relief to the Bank of England that there are signs the market is effecting its own natural slowdown without a rate rise, though underpinned by greater awareness among the public that the five-year holiday of record low interest rates is coming to an end,”

    Perhaps Carney has been cannier than we think-without actually having increased rates yet !

  25. ALEC

    You are right to remind me of that. At that time I was trying to remind people that Carney didn’t just choose unemployment -he said a range of indicators.

    But I have come to see that people focus on what they think is the key indicator-and grabbing a new one has just made things worse.

  26. JIM JAM thanks-I pretty much agree with that.

  27. @Colin
    “Perhaps Carney has been cannier than we think-without actually having increased rates yet !”

    Indeed. And even cannier by apparently persuading a large number of people to take his statements at face value.

  28. Colin – of course in Con/Lab marginals 2010 LDs will be targeted with a probably favourable break for Lab greater than UNS.

    I’m not saying anything new think most of us on this board feel that 2010 LDs and post 2010 UKIP support and how they break in marginals will be the main determinant of who gets most seats.

    Should there be a 2-3% direct Con/Lab or vice versa swing that would be meaningful as well but seems unlikley at the moment.

    (Scotland will have different dynamics of course whaterver the referendum result)

  29. Phil Haines

    I think you’re making a mountain out of a molehill with regards to Panelbase. When they released their statement last year:

    they said:

    There has been some discussion on Twitter of an organised signup of Yes campaigners, hoping to influence Panelbase polls. Not all of the quoted posts actually relate to political polls, but the point is still one to consider. In fact new joiners have had no significant effect on our results but we do recognise the potential for abuse of the system if people on either side were able to coordinate a mass sign up of new members. As a precaution, between now and the referendum we will not be including results from panel members who joined more recently than June 2013.

    So they haven’t closed their panel, they just don’t use anyone on it for Scottish political polls who joined after June 2013, presumably before the twitter fuss started. Panelbase had over 200,000 on their panel in April 2013 and have been going since 2004, so there will be some age on their panel as well. These will have joined to do consumer surveys, which is nearly all of what Panelbase does (they only do about one Scottish political poll a month or so[1], plus some non-VI opinion stuff in the North East). So any nationalist infiltrators will have been kept busy answering about washing powders and banks, just not independence.

    As discussed at the time, Panelbase’s poll with the odd question order was clearly perfectly valid private polling for the SNP which some idiot at their HQ decided to release on the grounds it looked good for them (which it didn’t because we all threw rotten tomatoes).

    As for Panelbase’s response they were technically correct because they didn’t split the sample to see how much difference the question order would make. It clearly did, but of course they said “we cannot really know for sure”. That’s what pollsters always say about everything. In part because it’s true; in part because someone might commission them to do some more work; in part because they can’t really say “Our client is a blithering idiot”.

    [1] Pro-rata I would reckon that a Scottish Panelbase member would be invited to a political poll every 2-3 years.

  30. JIM JAM


  31. @L Hamilton
    ” three polls from three different organistions in two days showing movement towards YES is worthy of note and probably real change.”

    A Yes/No poll can move in two directions. There are lots of polls.. Try flipping a coin twenty times. Did you get 3 heads in row, or 3 tails in a row? I rather think you did!

    Not noteworthy. Probably not a real change.

    And before you start searching for past posts of mine saying “the game’s up for Yes”, take my advice – you’re wasting your time. I’m objecting not because I’m anti-secession (though I freely admit I am), but because I hate this 3-in-a-row fairy story. It’s superstition. Lose it.

  32. @LHamilton

    “The momentum is now with YES, and while no doubt there will be further bumps, along the way it may well carry them to success over the next month.”

    Excluding DK’s Yes 43% (+3) No 57% (-3)

    That still looks suspiciously like a 14% lead for No to me, with less than four weeks to go before voting. Granted, the change is just a little outside MOE, so there appears to have been a slight movement towards Yes since the last poll, but this isn’t anything like the sort of momentum that will take Yes to victory.

    For that to happen, there now needs to be an absolute avalanche of opinion falling towards the Yes campaign over the last three weeks or so. Not just the vast majority of DKs plumping their way but significant numbers of those currently saying No to change their minds. It’s not going to happen.

    As for UK politics, the latest boringly “as you were” Populus, and the weekend YouGov, seems to point to last week being the polling equivalent of the Silly Season. Ranging from ICM showing a 7% Labour lead to an Ipsos/Mori tie, to a Populus Tory Lead, to an Opinium 21% for UKIP, it was all extraordinarily silly, wasn’t it?

    Back to normality thank goodness. Everybody calm down.


  33. I find extraordinarily silly that Yougov still fails to prompt for Ukip.

  34. Apologies if this Times Referendum poll from You Gov has already been posted:

  35. @Charles

    Apologies, my natural tendency to over-explain myself got the better of me there and you happened to be in the way.

    I did think Daodao’s original remarks (which drew me into the discussion) offensive – to paraphrase “You Christians like killing Jews so why not send ISIS after Israel, haha?”. Colin was wiser, and stepped elegantly aside.


    You are too kind-I actually did think he meant Hamas !!-should have known better !

  37. Charles
    That was some of the weird stuff I was referring to.
    Meanwhile , John Curtice on [email protected] downplaying any chance of a Yesss ! Victory.
    Worst poss outcome ? Narrow aye win.
    Best “. “. “. “. naw. “.
    If the second happens…..
    Post ref Lab offer Devo max, everything devolved except Defence and foreign policy. The West Lothian question is a shibboleth that can be cheerfully fudged if there is the (good) will on both sides.

  38. CROSSBAT11 and EWEN.
    I remain to be convinced about a Labour lead and the Lib Dems doing so well.

    Very unconvinced also about LVG.

  39. @Roger Mexico

    Thanks for that. So you’ve confirmed what I thought was the case but hadn’t seen any confirmation of – namely that the Panelbase panel remains closed from mid 2013 with regard to those used for political polling.

    A panel of 200,000 sounds an impressive number, but not when you consider that that is the total across the UK and we are only considering a Scottish sub-set of that for referendum polling. 20,000 perhaps, if pro-rata? Of those several thousand would need to be polled each time to get a sample of 1,000 depending on the response rate. And are people who respond to one survey asked more often subsequently, in an effort to get response rates up?

    I also think it’s reasonable to assume that panelists are more likely to be dormant with time so that in the absence of new entrants for political polling the response rate could be falling, which in turn raises risks of systematic differences in response rates.

    By all means give Panelbase the benefit of the doubt, but I do think you’re being too generous with regard to the order of their questions and their very grudging response when they were called out on that. It may be a case of the boy who cried “wolf” – that is, just because I’m less inclined to believe everything they say because of their past form and the commercial considerations that I referred to, it doesn’t mean that the wolf isn’t there now.

  40. Sorry that should have said Postage, not Charles.

    And yes, I like my Shibboleths with plenty of Fudge !

  41. CL

    LVG ? Luncheon Voucher Groupies ?

  42. Anyway Panelbase’s tables are now published:

    Two things strike me on the subsidiary questions. The first is about the NHS. Asked:

    Does the prospect of an increased role of the private sector in the NHS in England having an adverse effect on the Scottish budget which funds NHS Scotland make you likely or unlikely to vote for an independent Scotland in the referendum.

    most of the response is partisan, but 7% of No voters and 35% of Undecideds (v 5% less) say that it would be likely to make them vote Yes.

    Of course how plausible that scenario is, is another matter. But it ties in with the rising worries about the NHS we see in surveys across GB and also with Labour’s warnings elsewhere. Of course the real question would be more directed towards the continuation of the Barnett formula, but that’s another can of worms the No side might find difficult. And health is a particular problem in that even reassurances of continuation of the current situation will be met with the accusation that the Conservatives said there would be no top-down reorganisation of the (English) NHS and promptly embarked on one.

    The other interesting thing is about the question of trust. Panelbase asked Do you trust […], to stand up for Scotland’s interests? about a range of politicians. The results:

    ALEX SALMOND, First Minister 59-37 = +24

    NICOLA STURGEON, Deputy First Minister 53-33 = +20

    PATRICK HARVIE, leader of the Scottish Green Party 30-34 – -4

    DENNIS CANAVAN, Chairman of the Yes campaign 41-33 = +8

    JOHANN LAMONT, leader of Scottish Labour 30-40 = -10

    RUTH DAVIDSON, leader of the Scottish Conservatives 23-53 = -30

    WILLIE RENNIE, leader of the Scottish Liberal Democrats 20-43 = -23

    DAVID CAMERON, Prime Minister 19-66 = -47

    NICK CLEGG, Deputy Prime Minister 11-69 = -58

    ED MILIBAND, leader of the Labour Party 20-57 = -37

    ALISTAIR DARLING, Chairman of the No campaign 32-50 – -18

    Even allowing for partisan and nation bias, it’s still a pattern that weighs heavily against No. We saw this similarly in April when YouGov asked about which Party would best look after Scotland’s interests and the SNP got 45% with Labour second on 17%.

  43. Toonie

    Thanks for the link.

    I went to put the Holyrood figures into Scotland Votes, before realising it was pointless!

    1. It hasn’t been adjusted since Margo’s death, and reducing Ind to 0% on the list still gets her elected!

    2. Others at 7% on the constituency list is ridiculously high. The minor parties (excluding the LDs maybe) won’t put up a slate of candidates for constituencies. They stand on the regional list. Those Green & UKIP voters will have to select one of the main parties at constituency level (or not vote).

    However, now that YG includes place of birth for referendum purposes, it seems (not surprisingly) that there is a correlation (though less strong) between that and Holyrood VI.

    For the 3 main parties at constituency level, among those born –
    in Scotland – C 14% : Lab : 36% : SNP 41%
    other UK – C 16% : Lab : 44% : SNP19%
    outwith UK – C 23% : Lab : 28% : SNP19%

  44. @Ewen
    “I like my shibboleths with plenty of fudge”

    Shibboleth is in fact Hebrew for “ear of wheat” so you seem to be advertising a kosher beakfast cereal for kids.

    One for Dragons’ Den I think. I’d be in if I had any dosh as I’m in favour of fudge, both literal and metaphorical.

  45. @JimJam

    “I’m not saying anything new think most of us on this board feel that 2010 LDs and post 2010 UKIP support and how they break in marginals will be the main determinant of who gets most seats.”

    Agreed, except to add what happens to Greens where there is no candidate.

  46. New Thread

  47. Guymonde,

    Yes and how the LD hold up in their seats esp v Con will lefties still vote tactically in enough numbers.

    Both will impact but the big 2 are still UKIP current support and LD 2010 vote and how they will split in key seats.

  48. Has forward guidance become a ‘joke’?
    Carney has enabled himself to speak more openly and massage opinion perceptions and expectations.

    Forward guidance does not have to be a sledgehammer.

1 2 3 4