There are two Scottish polls in the Sunday papers, YouGov in the Scotland on Sunday and Progressive Scottish Opinion in the Mail on Sunday. Both show solid leads for the SNP in the constituency vote, and smaller SNP leads in the regional vote. Both also have the Green party in fourth place in the regional vote.

YouGov’s poll has the SNP lead falling, but that may well be just normal sample variation (as always, I’d advise caution in reading too much into movement of a few percentage points in a single poll until it becomes part of a trend). With just four days to go and the polls during the campaign showing a generally robust SNP lead, it would be surprising now if the SNP didn’t top the poll.

The YouGov topline figures with changes from 10 days ago are:

Holyrood constituency: CON 12%(+2), LAB 34%(+2), LDEM 7%(-1), SNP 42%(-3)
Holyrood regional: CON 12%(nc), LAB 33%(+4), LDEM 6%(-1), SNP 35%(-4), Green 7%(nc)

The Progressive poll’s topline figures with changes from a week ago are:

Holyrood constituency: CON 10%(+1), LAB 35%(-1), LDEM 6%(nc), SNP 45%(-1)
Holyrood regional: CON 8%(-2), LAB 36%(-1), LDEM 5%(-4), SNP 41%(+3), Green 6%(+1)


80 Responses to “YouGov and Progressive Scottish polls”

1 2
  1. FRANKG

    Correction to the multipliers in my post re NE Fife.

    In making some interface changes today in order to facilitate entry of polls in the standard SNP, Lab, Con, L-D order, I inadvertently swapped the 2007 actual %s. The multipliers should read:

        SNP = 42.0% / 33.02% = 1.27183 which normalises upwards to 10,529 votes from 6,169
        Lab = 34.0% / 32.39% = 1.04965 which normalises upwards to 3,257 votes from 2,312
        Con = 12.0% / 16.13% = 0.74410 which normalises downwards to 7,858 votes from 7,869
        L-D = 7.0% / 16.36% = 0.42792 which normalises downwards to 7,141 votes from 12,435

    As you can see the overall order doesn’t change because the 2007 Con & L-D vote shares were so close.

  2. Not been on for a while..
    Is there a guide to the Holyrood seats on here anywhere ?

  3. Following the correction noted in my previous post, the 3 YouGovs vs Scotland Votes are now:

    15th
        SV: SNP 55, Lab 48, Con 14, L-D 6, Grn 5, Ind 1
        BZ: SNP 52, Lab 49, Con 15, L-D 6, Grn 5, Ind 2

    21st
        SV: SNP 61, Lab 40, Con 13, L-D 6, Grn 8, Ind 1
        BZ: SNP 60, Lab 40, Con 13, L-D 7, Grn 7, Ind 2

    29th
        SV: SNP 54, Lab 47, Con 14, L-D 5, Grn 8, Ind 1
        BZ: SNP 53, Lab 48, Con 14, L-D 5, Grn 7, Ind 2

    The 5 L-D seats in the latest poll would be Tavish plus one regional seat in each of Highlands & Islands, Lothian, Scotland Mid & Fife and Scotland North East. Looks like my model is closer to SV’s than I thought.

  4. PAUL H-J
    Is there a guide to the Holyrood seats on here anywhere ?

    Not here that I’ve seen, but one of the best I’m aware of is STV’s at http://news.stv.tv/election-2011/battlegrounds/

  5. One link per post to avoid modding…..

    Scotland votes – http://www.scotlandvotes.com/ – also have reasonable profiles but DIY comments rather than analysis. A bit hit and miss but some are enlightening.

  6. This Scotsman interactive election map is quite useful too

    http://thesteamie.scotsman.com/election.aspx

  7. Rumuors of poll in the Scottish Sun Monday 2 May

    11% Constituency lead and 9% Regional Lead for SNP

    STV Poll due Tuesday

  8. Barbazenzero,

    Thank you. I tried the BBC which has a map/guide but nothing by way of real analysis on the seats.

    Would I be right in thinking that Cons could hold Ayr, Eastwood, Edinburgh Pentlands, Etrick / Roxburgh, and Galloway in Constituency section as a “good” result, with the balance coming from List MSPs ?

    For LDs, omens are not good, but most LD seats are more likely to fall to SNP, though they may have previously been LD/Con contests ?

  9. NATIONALIST
    Rumuors of poll in the Scottish Sun Monday 2 May …. STV Poll due Tuesday

    Any idea which pollsters?

  10. PAUL H-J

    My model with the latest YouGov shows the Cons holding only Ettrick, Roxburgh & Berwickshire plus 13 list seats.

  11. OldNat

    Bill Patricks comment about Tartan Tories is very much to the point of what I have found out about the thinking of Glasgow Labour voters.

    Few on here would quibble with the positioning of the parties calculated by by Scotland Voting Compass. It would be even more enlightening if they could produce a compass for all post-war elections and size the party emblems according to the share of the vote. We don’t doubt what that would show do we?

    I have long argued on these pages that Bill’s “traditional Unionist vote” was in fact a Christian Democrat vote and not recognised as such because these voters were Scottish and Presbyterian. To the metropolitan media both attributes belong on the image of Scotland which says “here be dragons.”

    These people are less numerous now because the influence of the churches is less, and the division between Presbyterian Unionist middle class and Catholic Labour working class no longer applies. Those that remain have moved over to the SNP, as Bill says, leaving the few “Powell/Thorneycroft Tories in a party dominated by English Nationalists and free-market fundamentalists who have marginalised the one-nation element and had never even recognised the existence of the Christian Democrat component simply because they were they were Presbyterian.

    These comfortaby off voters were repelled by the thought that they could be bribed by an unlooked for benefit in the poll tax, and were brought up to beleve that greed was bad and public service good.

    Bill is not quite right in saying that this has nothing to do with “that Tartan Tory nonsense”. That is so for us here or for anyone with a modicum of objectivity or ability to look at all the information together.

    The West of Scotland Labour Voter is different: neither well informed nor rational.

    The WSLV knows that the SNP grew from nothing at a time when the Conservatives declined and has been persuaded that these SNP politicians, supporters and voters must still be Tories at heart. There are only two possibilities arn’t there? “Our sort of people” and the evil Tories.

    We here know that is not so, and that the ex-Cons wandered loose from a party that did not acknowledge their values. Labour (Catholic and working class) was unappealing and unwelcoming, LibDems were a Highland regional party, and the SNP was a safe haven used because it was there, regardless of the SNP’s curious obsession with hypothetical constitutional matters.

    These people saw the SNP as party which understood traditional Scottish values of eglitarianism, community and a respect for education.

    The WSLV made up his/her mind which party to support in early elections as a young adult. Not the “nasty party” of Thatcher then. S/he hasn’t thought about it since.

    Therefore the WSLV has not noticed that the SNP and the Greens now occupy the place on Scottish Voting Compass that Labour did in their youth, and also does not recognise (or does not want to know) that NewLabour has moved closer to the Cons.

    To recognise either of these realities would be too disturbing,

    The WSLV will say that his or her parents and maybe grandparents voted Labour. Political opponents have misunderstood this as being offered a reason for voting Labour.

    It isn’t.

    It’s an expression of pride in the WSLV’s own unwavering constancy in doing the right thing for anything up to 60 years, and that parents did the right thing too, extending this pattern of good behaviour further into the past when the WSLV was too young to vote or not even born.

    The Labour party is an authoritarian cult (h ttp://members.shaw.ca/jeanaltemeyer/drbob/TheAuthoritarians.pdf) , It requires little effort and thought, and causes no coflict to be loyal to the party even if that means being disloyal to the principles the party once had and which attracted you to it in the first place. For the authoritarian follower (Caholic, working class and less educated) that is the easy option.

    Thus commited members of the Labour party (including one whose words and photograph appear on election literature) turn out to be Green or SNP on Scottish Voting Compass.

    If that analysis is correct, there is an opportunity for the SNP over the next five years to do substantial damage to SLAB at a time when the anti-Con element of Labour’s support may become as rootless as their colleagues in the Northern regions. If the two threats materialise over the same period of time, some party of the left must benefit. The Socialists are not in the game at the moment, so that leaves the Greens and the SNP.

    If the SNP cannot or does not attempt to, shatter the Tartan Tory myth, then the Greens have an opportunity to persuade WSLV that NewLabour is Tory-Lite and just as much to be avoided as the Tartan Tories.

    Conservatives can Bavarianise and re-brand and get back their lost sheep.

    In the d’Hondt Scottish parliament, the competition is not just betwen the big guns of the center parties, but betwen them and the snipers on the periphery who can bring about realignment.

  12. John B Dick

    As always, a thoughtful post. I’ll think about your analysis before I comment.

  13. Moraylooninbruseels,

    > Ian MacWhirter is now suggesting that Labour “borrows” one of the few SNP policies that they didn’t from day 1.

    What, independence? I always thought a Scottish Labour Party might get a good result… ;-)

  14. John B Dick,

    My point on Tartan Toryism was not that no-one believes it, but rather that I don’t think that the SNP’s success comes from being Tories in kilts. They’re not.

    As you suggest, the SNP has picked up the (culturally if not religiously) Presybeterian Christian Democratic voters who used to be the natural supporters of a One Nation, centrist Tory party. Most of the older SNP voters I know are former Tory voters, though this is largely because I live in what was once natural Tory territory.

    I find the West of Scotland Labour Voter just as difficult to understand as the West of Scotland Labour Party Member, simply because their political outlook is extremely different from mine. It would be even harder if I didn’t have the benefit of some time studying highschool history in Scotland, where UK history was largely the march of an expanding welfare state as politicians became more and more generous, followed by the nightmare of Thatcherism and finally ending happily the sunlight of New Labour (this was a fair few years ago now). I’m only marginally exaggerating the slant in the teaching.

    So I suppose I’m saying that I find the politics of Glasgow et al as opaque as the politics of Iran, because while there’s a lot of thinking going on there, it’s thinking from very different premises from my own.

    “These people saw the SNP as party which understood traditional Scottish values of eglitarianism, community and a respect for education.”

    I agree except for egalitarianism. Scottish culture, such as one can generalise, is not an egalitarian culture. Hence pining for the runrig system and the passion for absolute monarchs in Scottish history; the common concept of benevolent and expert elites in Presbyterianism and Scottish Catholicism; and the acceptance of severe inequalities in adjacent communities in most Scottish cities. It’s quite different from what I knew in either Ulster or England.

  15. @ Barbazenzero

    NE Fife

    Yep got your method of working it out from last time, but unfortunately having spent some considerable time, it still just doesn’t gel with me?

    For a starter you have not included ‘Others’ suitably multiplied up by their factor. For ‘Others’ your multiplication factor would seem to be 5.0/2.17 = multiplication factor of 2.3041.

    Next your 2007 notional votes total is wrong, it should read 29423. There were notional votes of 638 for Others.

    I am not quite understanding your rational for ‘normalisation’ by multiplying by 1.70 (corrected from 1.34 last time). I understand the principle in that what it does is to bring the total number of votes arrived at by your multification factors up to the total number of notional votes for 2007. (Accepting since you have neglected ‘Others’ from the multiplication process and also ‘Others’ from your 2007 Notional Total, your normalisation factor will have to change slightly.) But if your whole system is ‘correct’, why is the factored votes total not also the correct overall total? If your system was mathematically sound and logically then there should not be any necessity for a ‘normalisation’ factor.

    Incidently I got the Others % by adding up the %s shown in ‘Scotland Votes’ and finding it is 2.17% short. The number of votes this represents can then be calculated. Similarly with the Green/Others % on the Poll figures and for 2007 Notional overall figures

    I have noted your recent post re the differences between your system and SV. What would be interesting is to compare your Constituency party only totals with those of SV constituencies only. You can get that from the SV model by putting in the %s and taking away the regional seats total (and bars) shown above the regional prediction area. Since the method of calculating the regional allocation takes into account the constituency results, it has a sort of built in ‘error correction’. Thus significant differencies in constituency results had be ‘hidden’ within the overall totals. To get even small changes between your BZ and the SV that you highlight, could mean that there are still considerable differences in your constituency forecasts.

    Thank very much however for all your work and consideration in explaining your system and in dealing with my questions.

    Reference your forecasts for Cons. On my system from the constituency polls, I have them holding Ayr, Eastwood and also Ettrick/Roxburgh/Berwickshire

    I have Dumfrieshire as lost to the Lab and Galloway to the SNP. Edinburgh Pentlands I have as a toss-up between Con, SNP and Lab.

  16. John B Dick

    “The West of Scotland Labour Voter is different: neither well informed nor rational.”

    I’m always uncomfortable with theories that stereotype sections of the population.

    There is no single type of WSLV. I can recognise Labour voters like that, but I can do that for voters for every other party as well!

    Looking at the UK Political Compass, there is a cluster of parties which are Authoritarian and centre-right or right wing –

    BNP, Labour, DUP, UKIP, Tory.

    SLAB, IMO, are more authoritarian, locked in to Traditional Values, and centralist than their English colleagues. That appeals to a particular type of voter – and they’re welcome to them!

  17. Bearing in mind that YouGov’s panel selection for Scottish polls may be unreliable, the following comparison may have little or no value!

    However, I had a look at the change by social grade between a March poll (15-18/3) and the latest one, for the List vote – given the abandonment of the constituencies by a number of parties, the list may give a better picture of party support.

    Figures are for latest poll (changes since March in brackets)

    ABC1, C2DE, Party
    31(+3), 40(+5), SNP
    31 (-4), 34(-9), Lab
    16(+1) , 8(+1), Con
    8 (+1 ), 4 (-1) , LD
    8 (nc) , 7(+4), Green
    2 (-2) , 3 (-2) , SSP
    4 (+2), 3 (nc), Other

    Most changes are within moe, but the biggest shift from C2DE Labour to SNP and Green looks like it might be worth further investigation.

  18. FRANKG

    I didn’t import the 638 others from Prof Denver’s PDF because I picked the wrong source for the candidates and thought only “the big four” were standing. Now the BBC has all five (including UKIP) while SV has only “the big four” and STV oddly have four including UKIP but not the Con. It’s a shame that there’s no consolidated data source for this. But you’re right that I didn’t have them. Now I do, so thanks for that.

    I have attempted to exclude Other from notionals where only “the big four” are standing because the numbers are conjectural anyway. It is clear I haven’t double-checked my data there, however, and is something I will try to do PDQ.

    The national notionals I’m using are SNP 33.01%, Lab 32.38%, Con 16.12%, L-D 16.35% & Other 2.13%.

    Re “If your system was mathematically sound and logically then there should not be any necessity for a ‘normalisation’ factor.

    The multiplier applies to the national vote. So on this example the 33% who voted SNP have grown to 42%, so for every 100 who voted SNP in a plurality seat last time there will be 127 now. And for every 100 who voted L-D last time there will be 42 now. The multiplier for Other is indeed 2.34647

    But in any given area, the proportions of each party’s support will only exceptionally equal the national proportion. The multipliers should show the relative growth or shrinkage of support but the totals will seldom if ever equal the original number of notional votes cast in 2007. They can be normalised back to the notional votes by simple multiplication by the original total number of notional votes and division by the sum of all the multiples.

    The full calculation for Fife North East is based on 2007 notionals, now:
        Votes 29,423; SNP 6,169; Lab 2,312; Con 7,869; L-D 12,435; Other 638

    After applying the relevant multipliers I get:
        Units 22,953; SNP 7,848; Lab 2,428; Con 5,857; L-D 5,323; Other 1,497

    Normalising back to 29,423 votes, I get:
        SNP 10,061; Lab 3,112; Con 7,508; L-D 6,823; Other 1,919

    The normalistion process is only necessary to calculate national votes by party to check that they agree with the original input poll percentages, as the seats could equally be “called” based on the intermediate units. It’s the calling that is needed to provide initial plurality seat holdings to the d’Hondt process.

    Either way, and as before, the SNP take it with Con 2nd and L-D 3rd.

    I only started building this model just over a week ago, to try to satisfy myself that SV’s list seat predictions were meaningful, and although I’m still finding the odd bug – mainly as here due to bad data capture – either I must be doing something right or Scotland Votes are doing something very wrong, because my numbers are very close to theirs on most polls I have input so far. The difference in plurality vs list seats is also small.

  19. Additional finding from the YouGov poll being reported in the Scotsman

    “More people trust the SNP to defend Scotland’s interests when dealing with Westminster than Labour, according to a poll for The Scotsman.
    With just three days until polling day, the survey of 1,108 voters north of the Border shows 50 per cent believe the SNP is better at standing up for Scotland, while only 31 per cent think Labour is stronger.”

    It would be interesting to know if that judgment applies only to SLAB, or to UK Labour as a whole.

  20. @Barbazenzero

    Thanks again for the clear explanation. I did understand the mechanics of your system, it’s the logic for the ‘normalisation’ part that I have trouble with justifying to myself. If your system was totally viable then such a ‘normalisation’ process should not be necessary.

    With regards to your BZ comparision with SV, I have done the SV calc for you for the constituency seats and the recent Yougov poll.

    Constituency Poll
    Con 12
    Lab 34
    LD 7
    SNP 42
    Others 5

    The SV predictor gives the constituency allocation as follows:
    Lab 26
    SNP 40
    LD 4
    Con 3
    Grn 0
    Ind 0

    Presumably the LD 4 seats are Edinburgh Southern, NE Fife, Orkneys, Shetlands.
    Presumably the 3 Con seats are Ayr, Eastwood, Ett/Rox/Berw

    What does your system have as the constituency seat holdings and which are the LDs and Cons constituency seats?

    Please may it make very clear – I am not saying that SV system is any more or less valid than your own. Merely that for the constituency seats you seem to be significantly different. Because the Regional List allocation takes into account Constituency holdings, then it should have the effect of cancelling most minor differencies between your systems in your overall final totals. If you are still comiing up with overall differences then you might actually have significant differences in constituency holdings. From your previous comments on LD and Con constituency seats held it would appear that significant differences do exist in your forecasts.

  21. Oldnat @ John B Dick

    “The West of Scotland Labour Voter is different: neither well informed nor rational.”

    …… . I can recognise Labour voters like that, but I can do that for voters for every other party as well!”

    My recent discovery was that WSLV’s I have spoken to actually believe the TARTAN TORY misrepresentation.

    Nobody else does.

    Bill Patrick

    As I have said before the SNP’s USP is competence pragmatism and rural sparsity practical solutions. In the Northern regions their claim to be the best lodging for the anti-Con gives them a current advantage over the LibDems for obvious reasons.

    The Central Belt anti-Con prefers Labour. Even the preference for Lab in the UK election and SNP for the Scottish one is an anti-Con choice

    Egalitarianism? The Reformation – The national poet – Education – (Private Schools) – The NHS (Higher 1947 funding is still there. Sparsity and morbdity are excuses, not the reason) – history of co-ops and mutuals …

    The Scottish Christian Democrats were a generation ahead of English Conservatives on colonies; race; women clergy; gay clergy as well as other things.

  22. For what it is worth, using the average of the two polls and uniform national swing my spreadsheet gives SNP 56 (42+14), Labour 51 (26+25), Tories 10 (2+8), LDs 5 (3+2) and Greens 7.

    Using multipliers it gives SNP 58 (43+15), Labour 49 (28+21), Tories 12 (1+11), LDs 4 (1+3) and Greens 6. SNP constituency gains include Airdrie, Cumbernauld, East Kilbride, Falkirk East, G Cathcart, G Kelvin, Orkney, every mainland seat they didn’t hold north of Cowdenbeath, Pentlands, E Western, Midlothian North, Ayr, Clydesdale, Galloway and Dumbarton. (It doesn’t include East Lothian, though this would fall if you assume uniform swing.)

    One of the questions that has to be asked is whether the SNP is actually campaigning in all these constituencies given how far they were behind in some of them in 2007. On the other hand, in AMS election you always have to campaign everywhere anyway, so I guess the answer is probably yes.

    Also, note that the SNP constituency gains means that the list vote is now likely to be the determining factor in all regions except NE – Labour could well gain list seats in Glasgow or Central. This also means that the last seat is often very close.

  23. CHRISTIAN SCHMIDT [& FRANKG]

    using … uniform national … gives SNP 56 (42+14), Labour 51 (26+25), Tories 10 (2+8), LDs 5 (3+2) and Greens 7.
    Using multipliers it gives SNP 58 (43+15), Labour 49 (28+21), Tories 12 (1+11), LDs 4 (1+3) and Greens 6.

    We’re getting similar results here. I certainly don’t claim one method to be “better” than the other. What I’m most interested in is how each model compares when we have the actual seat counts as well as the actual national percentages, to see which is closer. I’ll post more details re FRANKG’s mathematical concerns next week-end if the multiplier approach “wins”.

    Also, note that the SNP constituency gains means that the list vote is now likely to be the determining factor in all regions except NE – Labour could well gain list seats in Glasgow or Central. This also means that the last seat is often very close.

    Aboslutely, and especially if the list vote is closer between the big two on the list than in the plurality seats. It would certainly be good to see at least one more “repectable” poll before Thursday which is free from the worries over unreliability in the proportions of “Holyrood loyal” and “Holyrood disloyal” in YouGov’s Labour political ID.

    The continuing unpopularity of Iain Gray for FM even within his own party may even have spawned a third category of “Holyrood constituency loyal but list disloyal” containing a mix of those planning to vote other than for Labour on the list. That could ultimately prove more directly damaging to Labour’s seat count than the fall in the plurality vote.

  24. Just saw this tweet from
    Scotlandvotes Weber Shandwick
    Very interesting polling numbers on the way tonight! #sp11

    No other info, I’m afraid, I guess they’re trying to excite the crowds…

  25. Nicola Sturgeon (SNP deputy leader for non-Scots) has just re-tweeted this from STV’s Jamie Livingstone

    “Dramatic new poll for STV – tune in at six. #sp11?

    I assume Nicola would not have retweeted bad news but that’s pure speculation on my behalf… ;-)

  26. I read the link, no details on who conducted it and when.

    A note: I just completed a very long poll for Yougov, asking a host of questions, not just voting intentions in Scotland. It had independence, Alternative Vote, views of the Lib Dems, Scottish Parliament, and the Banking crisis.

  27. Interesting poll, ignoring silly party comments.

    An SNP Green majority would be interesting.

  28. Very interesting poll and it is nice to see one that (I hope!) more realistically illustrates Conservative support. I can’t see the disparity between Labour and the SNP being nearly as great on the day though…

  29. STV Poll

    Constituency
    SNP: 45%
    Lab: 27%
    Con: 15%
    LD: 10%

    Regional
    SNP: 38%
    Lab: 25%
    Con: 16%
    LD: 9%
    Green: 8%

    Seats (via http://scotlandvotes.com)

    SNP: 61 (+14)
    Lab: 32 (-14)
    Con: 18 (+1)
    LD: 9 (-7)
    Green: 8 (+6)

    Cons gain 1? Big 2 gap of 1 increased to 29? Parliamentary Majority for independence? SNP three short of an overall majority (Lab PO)? No Margo?

    None of these things is likely.

    Voodoo poll (or at least prediction.

1 2