Unless proximity to the election has cancelled out reluctance to poll over a bank holiday weekend we won’t be getting any new polls for a day or two. In the meantime I should really update with a Scottish poll that came out over the weekend.

The Scottish Opinion poll for the Scottish Mail on Sunday had Westminster voting intentions of CON 11%, LAB 27%, LDEM 11%, SNP 43%. The poll had a sample size of only 650, so is what we’d normally consider just a snapshot poll in GB terms, and was conducted between 15th and 20th May.

The poll is rather out of line with other Scottish Westminster voting intention polls (though ironically, is very close to what other companies have for Holyrood intentions) – the Conservatives are far lower than in other pollas, Labour lower and the SNP much higher. There is no recent track record of Westminster polls for Scottish Opinion to look or, nor as far as I am aware are they members of the BPC, so we can’t look at their details tables for an idea of where the differences come from.

31 Responses to “Catching up with the weekend’s Scottish polling”

  1. any polling data coming out tonight, and yes this poll seams a bit out not in line with other polls

  2. I am inclined to dismiss this poll due to its rather unrepresenative size, and the fact its entirely out of line with other Scottish opinion polls.

    This seems to me a blip originating from an unreliable sample (in both size and weighting).

  3. This Poll seems to be in about the right area for the Labour share maybe 1 or 2 % too high but on par !

  4. This may not look in line with recent Scottish Polls but it certainly looks in line with several more recent UK wide polls that put the SNP on 4% of the UK vote.

  5. Easy to dismiss this as a small sample, but to ignore the fact that its numbers are also on par with those shown in Scottish subsets in larger UK polls is questionable.

    The next full Scottish poll will tell the tale.

  6. “Easy to dismiss this as a small sample, but to ignore the fact that its numbers are also on par with those shown in Scottish subsets in larger UK polls is questionable”

    Well that doesn’t say anything at all, because the subsets were too small as well.

    This is clearly unrepresentative of Scottish voting intentions.

  7. This is far too small, with an unpublished methodology.

    Best to ignore it, really.


    4% is far too small to extrapolate to Scotland; simply given rounding and a standard error of +-2%, the SNP’s vote could be between 18% and 78% in Scotland.

  8. I could not say if the Scottish Opinion indications are completely accurate.

    They have a mixed record as a polling company although they have been on the go for a number of years and certainly have no track record of bias towards the SNP.

    I think origonally (but not now) they were part of the Daily Record marketing operation and have been usually considered kind to Labour.

    However it is true that these remarkable figures are supported by the recent indications from the Scottish samples of UK polls which are small but are all showing the same pattern of an SNP surge.

    Neil and Dean should not dismiss this poll out of hand. The Euros may show that it is closer than you think.

  9. Seen in isolation not relevant but does fit the tendencies of other small sets…

    And if I was the SNP I’d still be happy….

  10. It’s a small sample, but not a vary small sample: a sample size of 650 gives a “margin of error” of 4% so sampling error is insufficient to account for the large shift from Con+Lab to SNP. It would be nice if we had the tables and details of their methodology but there is no a priori reason to suppose that their methods are flawed and there is an obvious political explanation for the results, so it should be taken seriously.

  11. the last sunday times poll had the tories on 28%,for the third time(not widely reported on this site presumably because it showed the tories in a good light)so i think this is incorrect.

  12. I have to say that Jamie Smith has it dead on, plus even the usually unflattering system three TNS polling places the tories on 19% in Scotland- and all YouGov polling for Scotland only places them on 20% (except for the ‘pinch of salt’ UK wide scottish sample placing the tories on 28%).

    My point is tha unless a massive and inexplicable shift has occured this is hardly a poll worth conserning oneslf with. Especially since the Scottish EUROS outcome is most likely going to be 2:2:2- which hardly supports tories on 11%!


    The liberal level of support does reflect recent trends (placing them between 11-14%), and Labour on 27% should be their level of support if the UK trend is reciprocated in Scotland; but as we all know (or at least ought to) Scotland never follows national (UK) trends. If Labour fell to 27% I’d get horrendously drunk in celebration! (sorry….being bias, couldnt resist).

  13. This poll seems to be accurate and the Labour figure of 27% is IMO about 2% too high. I think they will be in the range 20-25% in Scotland come polling day

  14. I too have my doubts about this poll and particularly the Tories on 11%, but I don’t doubt that it indicates the current mood in Scotland.

    The expenses affair has hurt all three main parties more than smaller ones, and the `SNP by a combination of fewer MP’s and it being a westminster scandal is best placed to benefit.

    Although nationally it is Labour that has come off worst and Cameron has played it far better than Brown, in Scottish terms it is quite possible that more people have reacted to a long held anti Tory bias than elsewhere.

    Equally strong in some cases almost tribal support for Labour might have mean’t that they have got off lighter than in the UK as a whole.

    11% is a big fall from the last two full Scottish polls of 20%, but for most of the year the Tories have averaged about 17%. I wouldn’t in any way rule out the current expenses scandal pushing them back down to that level ,reversing recent gains. I wouldn’t however accept anything below 16%.

    As for Labour although they have been in the high 30’s recnetly again in the Brown slump last year they did sink below 30%, so although 27% would be a new low given recent coverage dropping below last years 29% isn’t unreasonable.

    As for the SNP or highest ever has been below 40% so even with the benfit of gains by all Parties I wouldn’t see us above that.

    As a 650 sample has a 4% margin of error then we could see this as;

    Labour 31%, Tory 15%, Libdem 14%, SNP 39%.

    I’d hazrd as guess that given a larger share share for independents next week the Scottish result will be close to;

    Lab 25%, Tory 15%, LibDem 10%, SNP 35%, Grn 5%, UKIP 5%, Others 5%.


  15. I think 14% is the upper most maximum for the Liberals in Scotland; based on all recent polling if they managed 14% they’d be making a signifcant recovery.

    Labour on 25%? Nope; just cant see it myself, the lowest Labour will get I think in 29%- even then thats a major push into the core tribal labour vote, and such labour lows ought to see a major push upwards in the SNP share. (However I would point out that in the last YouGov it saw all parties and no just the SNP benefiting from the Labour slump down to 32%).

    But sorry, 15% for the Tories, jsut after having a major conference in Perth, and D.C taking the lead in the expenses scandal? Not a chance, I’ll make a bet with anyone who can honestly believe what Peter is saying- the tories on 15%. Utter hope (if your not a scottish tory that is).

    My reckoning for the next poll? It will show something in the region of:

    SNP: 32%
    Lab: 29%
    Con: 22%
    lib: 14%

    (I’m basing my averages on YouGov Scotland-wide polling, excludes the TNS system three polling, I tend to find that it unfairly underestimates tory support levels).

  16. Dean,

    I’d be very surprised if the Tories get 22% in the next full Scottish poll . 20% is more likely and perhaps 18%. Regardless of poll averages over recent months the public are furious over the expenses affair and it’s labour and the Tories who have come off worst.

    In the UK with Brown looking clumsy and Cameron sure footed Labour will come off worst, but in Scotland with a Scottish PM and disgruntled Labour voters unlikely to switch straight to the Tories having the Tories stay at the high point of their vote is over optimistic.

    So for the next Scottish poll, if it comes in the next fortnight and it could be in three days, will have the Tories below 22%.

    As to next week I am sticking with what I said, in the EU elections next Thursday the Tories will get about 15%.

    In 2004 the shares were;

    Lab 26%, Tory 18%, LibDem 13%, SNP 20%, Grn 7%, UKIP 7%, SSP 5%, Others 4%.

    Given expenses the Tories down to 15% is in the ball park.


  17. Whens the next poll ? Hopefully its not ICM because they are useless

  18. The Conservative and ScotLibDem are unfeasably low as is the extent of growth in SNP support but there are undoubtedly trends in the directions indicated, albeit less dramatic than this result.

    The one clear inference with a plainly obvious explanation is that in four party Scotland the negative voter who wishes to vote against the government of the day has greater choice than in England and has no need to vote for the main opposition party or even the “wasted vote” LibDems.

    In current circumstances, since the Conservatives are also implicated in the expenses scandal they are no position to overcome the unwillingness of most Scots to vote for the party of Thatcher. If the SNP did not exist either the LibDems or the Greens would be the beneficiaries of Labour unpopularity.

    As things stand, “the time has come” for the SNP and it is not of their doing or caused by an appreciation of their arguments in favour of independence. Rather it is the case that independence may follow the success of the SNP.

    If so, the many on thesepages who are of the opinion that elections are lost, not won, will understand. Unionists who believe party propoganda will be confused and the nationalists will never believe that they have not won the argument.

    If you are looking for a benchmark for the limits to which a party of government can lose support against which you can compare the party which in 1951 got more than half the votes cast in Scotland, then consider the renamed East German Communists who managed 17.1% in free elections in the former DDR states within the reunified Germany.

    Compare that with the 11% above for Scottish Conservatives. Can it really be quite as bad as that? Surely not, but it doesn’t need to be for them to gain hardly any more seats in addition to the two or three at most expected a year ago.

    Before the banking and expenses issues I was of the opinion that we were on a slippery slope to independence. Now, I think the time is near when the Unionist parties will give up trying to prevent the inevitable just as the DDR Communists did.

  19. We will find out at the GE whether he expectation that the Conservatives will form a government is enough to overcome the reluctnce of Scottish Conservative voters to tell pollsters the truth about their intentions.

  20. @John B Dick

    The DDR’s SED was never elected in free and fair elections, and never achieved more than 20% in its formerly DDR states since such elections were introduced.

    The unionist parties in Scotland command approximately half of the vote in Scotland even in this low-number, unknown-methodology poll in the middle of a constitutional crisis.

    Also, since independence is such a massive act, nationalist parties would need to show dominance in the style of the Irish Parliamentary Party in the start of the 20th century, rather than “about half” as seen in Catalunya, Spain or Quebec, Canada.

  21. I have heard through PB that we are due a YouGov survey for the Telegraph tonight and Comres and Mori are in the field at the moment so expect these at the weekend. It looks like we wont be getting ICM who are useless anyway !

  22. I may regret this, but go on then. Give me a reasoned critique of ICM’s methodology.

  23. Anthony,

    It is quite simple – they tend to allocate a larger amount of don’t knows that favours the lower polling party ie. Labour ! It doesn’t take rocket science to determine that they are way out of line with the other major pollsters – just look at your table from the last 4 months! In that respect I consider them useless, I am entitled to my opinion!

  24. Yes, they do, but what’s wrong with that? The re-allocation of don’t knows is supported by re-contact surveys after the election – ICM and Populus have both re-contacted people after general elections, and about 50% of them do indeed vote for the party they say they voted for last time. Their methodology is evidence based. Equally, ICM have been using this method for 15 years, and it has a very strong track record of success.

    Being out of line with others doesn’t mean you are wrong (and for the record, ICM aren’t out of line with Populus). ICM were also out of line in 1997 – they were right, and everyone else was wrong. Harris were out of line with everyone else in 1992 – they were right, and everyone else was wrong.

    Besides, the re-allocation of don’t knows isn’t the reason for the difference anyway. Without the re-allocation, ICM’s last poll would still only have been showing an 12 point lead.

  25. Okay Anthony only results from the big day next week and whenever the GE is will tell us who are the most accurate !!

  26. Wayne,

    “Okay Anthony only results from the big day next week and whenever the GE is will tell us who are the most accurate !!”

    Not necessarily.

    The methodology that best shows how people think they would vote if there was a vote now, doesn’t necessarily show how they will vote in a real election in the future.

    ICM could be wrong now but not in an election or vica versa, we just can’t tell.


  27. Indeed – not necessarily. Different methodologies could cope very well with a general election (high turnout, low impact from minor parties), but not cope well with European elections (low turnout, high impact from minor parties).

  28. Curiouser and curiouser…..


  29. ICM have the BEST track record of all pollsters when it comes to getting the final result most accurately. ICM are, infact, the gold standard of pollsters.

    However, when it comes to local elections, euro elections, by elections, etc… they do of course strggle much more, but then so do all the pollsters.

  30. I take your point John, and Peter about the expenses damage, and it very well could have the adverse effect that you outline- damaging the tories more than labour.

    But from the UK polling, its labour who have fallen to the depths of 22%, and not th tories. But in th Scottish case, I just havent felt that ‘anti-tory’ feeling when out canvasing and leafleting, even when it was Bannockburn I was in! (Oh and damn good work must go out to the SNP, the Labour candidate only just scraped home on 60 votes!). But if labour support gets a hit, then its only natural that it shall fragment, with a good 2/3 going straight to the SNP, sure- but that last third I’d expect to splinter between liberals and tory (both of which have taken the lead in tackling the abuse by MPs) and that matched with the favourable feelings for auntie bella, and the perth conference message ought to hold us well above 15%.

    It will be the euros that shall tell all (with a pinch of salt due to the usually low turnout)- but I’ll bet money (and have!) that it will be 2:2:2- and that ought to confirm the strenght of the tory vote surely?

  31. Dean,

    You need to separate out Westmister intentions – which will be for a GE either this autumn or in May 2010 – and likely result in the Euros – now less than a week away.

    You can probably take at least 2-4% off Westminster share for each of the four main parties to arrive at likely figures for Euros since others (of which there are 9 choices) will take at least 20% if not 30%.

    I disagree with Peter’s projection above for next week as I think he has SNP and Lab too high and Tories and Others too low – the latter much too low.

    In my view we will probably see:

    SNP – 27-30%
    Lab – 20-23%
    Con – 18-21%
    LD – 9-11%
    Others (incl Green Ukip etc) – 22-28%

    The seats are still most likely to split 2:2:2 but 2:2:1:1 or even 3:2:1 may happen.