Fifteen Weeks to go

Week three of the year and the regular cycle of opinion polling is back to full speed, with the first ComRes and ICM polls of the year. Almost all the regular polling companies have now reported figures from 2015 (we’re only waiting for Survation and ComRes’s telephone series).

ComRes/Independent on Sun. (15/1/15) – CON 33%, LAB 34%, LDEM 7%, UKIP 18%, GRN 3%
YouGov/Sun on Sun. (15/1/15) – CON 31%, LAB 33%, LDEM 7%, UKIP 16%, GRN 7%
Populus (15/1/15) – CON 32%, LAB 35%, LDEM 9%, UKIP 14%, GRN 6%
YouGov/S Times (16/1/15) – CON 31%, LAB 32%, LDEM 7%, UKIP 18%, GRN 7%
Opinium/Observer (16/1/15) – CON 28%, LAB 33%, LDEM 7%, UKIP 20%, GRN 6%
Ashcroft (18/1/15) – CON 29%, LAB 28%, LDEM 9%, UKIP 15%, GRN 11%
Populus (18/1/15) – CON 35%, LAB 36%, LDEM 8%, UKIP 13%, GRN 4%
TNS (19/1/15) – CON 31%, LAB 31%, LDEM 8%, UKIP 16%, GRN 7%
YouGov/Sun (19/1/15) – CON 32%, LAB 32%, LDEM 8%, UKIP 15%, GRN 7%
ICM/Guardian (19/1/15) – CON 30%, LAB 33%, LDEM 11%, UKIP 11%, GRN 9%
YouGov/Sun (20/1/15) – CON 32%, LAB 30%, LDEM 8%, UKIP 15%, GRN 10%
YouGov/Sun (21/1/15) – CON 33%, LAB 34%, LDEM 6%, UKIP 14%, GRN 8%
YouGov/Sun (22/1/15) – CON 31%, LAB 33%, LDEM 7%, UKIP 17%, GRN 8%
Populus (22/1/15) – CON 32%, LAB 36%, LDEM 9%, UKIP 13%, GRN 6%

The horse race remains extremely tight between Labour and the Conservatives, with most polls showing them within a point or two of each other, generally with Labour marginally ahead of the Tories. The UKPR average now stands at CON 32%(-1), LAB 33%(nc), LDEM 8%(+1), UKIP 15%(nc), GRN 7%(nc). The Green party’s rising support got a lot of attention this week because of their double digit figures from Ashcroft and YouGov, but there has not been a sudden step change in their level of support, it’s been on a steady increase since last year.

Scottish polling

This week we also had two Scottish polls from Ipsos MORI and Survation. Both continued to show a solid lead for the SNP in Westminster voting intentions. MORI had SNP on 52%(unchanged) to Labour’s 24%(up one). Survation had the SNP on 46% (down two) and Labour on 26% (up two). Both would still translate into an SNP landslide in Scotland come May.

Week Three

Every general election seems to start with the political parties putting out a flurry of announcements at the start of January, and then running out of steam a bit. This week’s political news has been rather bitty.

  • The reporting of the Chilcot Inquiry has been put back until after the general election, we can expect to see some polling on that at the weekend.
  • Peter Mandleson criticised his own party’s mansion tax proposals. Nationwide the idea of a mansion tax has extremely wide support – in September YouGov found 72% support for a tax on properties over £2million pounds. Criticism of it from within Labour tends to come from London, where it is less overwhelmingly popular, but still gets the thumbs up – YouGov London polling last August found 49% of people in London supported a “mansion tax”, 18% were opposed. Amongst London’s Labour voters 61% supported the idea.
  • The government announced that they would after all introduce legislation on plain packaging for cigarettes before the election, something that has previously been in and out of the long grass, and was seen as one of those policies that the Conservatives had put away as part of “cleaning the barnacles from the boat”. Generally speaking there is public support for the proposal – YouGov polling for the Sunday Times in July in 2013 found 58% of people supported compulsory plain packs, 26% were opposed. YouGov polling for Ash in 2014 that included a picture of an example of a plain pack found 66% of people in support, 10% opposed.
  • Finally the debates debate rumbles on, with the broadcasters making a new proposal to include the Greens in the debates… but also to include the SNP and Plaid, so that the format becomes two debates between seven leaders, and one debate between just Cameron and Miliband. Including all seven leaders was actually the most popular single option in the YouGov/Sun on Sunday polling last weekend, chosen by 35% of people. Between them though 49% of people preferred one or another of the options including fewer leaders.


The latest forecasts from Election Forecast, May 2015 and Elections Etc are below. All are still predicting a hung Parliament. Note that Steve Fisher has made some substantial changes to his Elections Etc model in order to treat England and Scotland separately, and hence reflect the increase in SNP support in Scotland

Elections Etc – Hung Parliament, CON 283(nc), LAB 278(-3), LD 23(-3), SNP 41(+5), UKIP 3(nc)
Election Forecast – Hung Parliament, CON 278(+1), LAB 286(-3), LD 28(+1), SNP 34(+2), UKIP 3(nc)
May 2015 – Hung Parliament, CON 269(-4), LAB 289(+9), LD 27(+3), SNP 38(-8), UKIP 4(nc)

270 Responses to “Fifteen Weeks to go”

1 4 5 6

    “A leaked general election strategy paper shows that activists in fourteen seats, including key SNP targets in Glasgow and Lanarkshire, have been in touch with fewer than 100 voters”

    They couldn’t even get enough support to go ahead with a Burns supper in Edinburgh last week and it even came with a speech from the star attraction herself………kezia dugdale!!

  2. UKIP defecting back to the mother party…….It’s going to be a funny ole Sunday.

  3. Apologies for length of this post!

    Really is beginning to look like the Lib Dem vote in Scotland will be very significantly lower in 2015 than previously.

    Without any Ashcroft constituency figures I have tended to assume that popular Lib Dem incumbents (Carmichael, Kennedy, Moore, Thurso) will survive and that others (Alexander, Crockart, Smith, Swinson,) have a fighting chance.

    But are there actually enough Lib Dem voters for any of them to survive?

    If we take the 5% figure from the YouGov crossbreaks as accurate and apply that to the total 2010 Scottish votes of around 2.5m that gives us around 125,000 Lib Dem voters in the whole country.

    Even If we apply some optimistic figures for Lib Dems and give them their best post independence vote poll figure of 7% (Survation, 16 Jan) that is still only 175,000 voters.

    In 2010 there were 42 Scottish Constituencies where the Lib Dems came 3rd or 4th with between 5.0% (Glasgow East) and 19.4% (Edinburgh East) of the vote. Average is around 10%.

    Assuming a UNS reduction of 75% in the Lib Dem vote these will fall to between 1.5% and 5% with an average of around 3%.

    As a best case scenario for Lib Dems let’s call this 2% and that will translate to around 35,000 Lib Dem voters in these hopeless constituencies.

    Then we have 6 Constituencies from 2010 where the Lib Dems came 2nd:

    15.8% Glasgow North West
    28.4% Aberdeen South
    31.3% Glasgow North
    33.8% Edinburgh North & Leith
    34.0% Edinburgh South
    35.1% Dunfermline & West Fife

    This is a total of 75,000 votes in 2010 which on UNS would reduce to around 19,000 votes.

    Now the good news for the Lib Dems here may be that these seats look like they might be even worse for them than the national average. These are seats with lots of students (Ed South, Glasgow N and NW, Aberdeen South) or ex Lab protest voters (Dunfermline, Ed North).

    The Lib Dem vote could completely collapse here leaving more voters in Lib Dem rural heartlands. Let’s assume the Lib Dem collapse is twice the national average here and they drop to 10,000 votes across these 6 seats.

    So we have 45,000 Lib Dem votes across 48 Constituencies leaving us 80,000 – 130,000 Lib Dem votes across the 11 Lib Dem held seats.

    I’m going to divide these into:
    Strong Incumbent (Orkney & Shetland, Ross, Skye & Lochaber, Caithness, Sutherland & Easter Ross and Berwickshire, Roxburgh & Selkirk.)
    64,000 2010 Lib Dem votes

    Weak Incumbent (Argyll & Bute, East Dunbartonshire, Edinburgh West, Inverness, Nairn, Badenoch & Strathspey and West Aberdeenshire & Kincardine.)
    86,000 2010 Lib Dem votes

    No incumbent (Gordon, North East Fife).
    35,000 2010 Lib Dem votes

    Now I’ll assume the Strong Incumbents perform much better than the Lib Dem average and only drop a third of their 2010 vote, enough for them all to hang on with reduced majorities.

    This uses up 40,000 or so 2015 voters leaving us with 40,000 – 90,000 2015 voters to distribute across the remaining 7 seats.

    Even at the upper end of this 13,000 votes per seat is only likely to be enough in smaller seats or those with even 4 way splits of the vote. Realistically the Lib Dems would need to rely on losing more votes than UNS in seats like East Dunbartonshire and Gordon to leave them in with a chance in seats like Inverness and Argyll & Bute.

    Overall my figures suggest that even reduced to 5-7% of the Scottish Vote it may be possible for Liberal Democrats to retain 4-6 seats rather than the 1-2 seats suggested by UNS.

    This would need them to collapse in the central belt amongst Ex Lab anti Iraq War types and lose all their Edinburgh/Glasgow/Aberdeen student voters while holding up their vote in rural areas with popular incumbents.

    Anecdotally I think this is not an unlikely scenario but would like to see some Ashcroft polling to support it.

    My predictions

    Ashcroft polling shows Lib Dems to hold:
    Orkney & Shetland,
    Ross, Skye & Lochaber,
    Caithness, Sutherland & Easter Ross
    Berwickshire, Roxburgh & Selkirk

    Ashcroft polling shows close fights with Lib Dems in contention:
    Inverness, Nairn, Badenoch & Strathspey
    Argyll & Bute
    North East Fife

    Ashcroft polling shows Lib Dems well behind in:
    East Dunbartonshire,
    Edinburgh West
    West Aberdeenshire & Kincardine

  4. Re my previous ,approval on economy plus 5,dc plus 6 ,family standard of living improving in next 12 months plus 7.

    Sooner or later this will translate into VI.


    Great analysis. Gordon has gone I’m sure and it will be one of the more interesting contests for obvious reasons.

    If the Scottish 2011 election is anything to go by where the Lib’s lost all of their mainland seats on 7.9% then I think the prospect of them holding onto more than two seats in May looks rather grim but local factors and tactical voting by toffee nose lairds could keep them some of their seats but personally I think the Lib/Dems are finished and and at best hold onto their Borders seat and O&S.

  6. Here is this week’s YG VI by 2010 voter ID:

    I have made a small adjustment to the change graph. There were lots of small data points, and given the usual churn etc, were probably not significant.

    To add a significance filter, I have calculated for each group (2010 Con, Lab, LD and Other) the standard error (standard deviation / square root (5)) * 1.96. Data points within this value have been excluded. Please note the square root of 5 is based on 5 YG samples for the week, 1.96 has been used for 95% CI.

  7. ” Please note the square root of 5 is based on 5 YG samples for the week, 1.96 has been used for 95% CI.”

    Stuff like that should be preceded by a warning: my head hurts now.

    I left school precisely to escape such language.


    @” the population growth is directly correlated with education, urbanisation and related awareness of and access to contraception.”

    From the research I read, that is a significant part. I don’t think there is a conflict between this & NeilA’s” Religious/cultural practice, &economic insecurity”

    Religion can mitigate against birth control.
    Children are a substitute for welfare & social care systems in poor countries.
    Subsistence family farming must substitute assured manpower supply for lack of technology.

    For me the common denominator in all of this is the role & place of women.
    Where women are not constrained by culture & circumstance , and are free to choose the number of children they conceive, population increase falls.

    that is very interesting analysis thanks john

  10. COLIN

    One of your comments has evaporated. Best not discuss Mr Bean.


  11. Charts updated folks. Had to adjust the UK 15-poll averages down to allow for both Con and Lab drops (a sign of prolonged lower averages). Both now teeter on averages around 35% in Mids & Wales too.

    You can get a good idea of the more recent Con / Lab changes in the UK 30-poll chart. The first three data points versus the last three show how the gap between them has narrowed. Then you can look at the regional 30-poll charts to see where it all happened (not just Scotland).

    h ttp://

    The calendar charts similar data to the 30-poll charts (30 polls amount to 6 weeks), but it’s not quite as visually informative, and indeed can make the viewer see more change than is happening (see RoS Conservative calendar v 30-poll).

    h ttp://

    Obviously the 3-month change charts show you the real movement over the period, but it is a snapshot between two polls, rather than between two averages, so treat with caution.

    Government Approval for 2015 so far:

    UK – Up slightly on December
    London – Little change
    RoS – Up and getting close to positive ratings
    Mids & Wales – Little change
    North – Up slightly
    Scotland – Up, but at minus 41%, not exactly stunning.

    h ttp://

    Leadership Ratings:

    Cameron – Edging towards positive ratings in the UK as a whole as he hits minus 7 points. Already well there in RoS and just nudged into positive ratings in London. Still well into negative ratings elsewhere.

    Miliband / Clegg – Very similar ratings. A race to the bottom, it seems. The only big difference is that Miliband is at -40 in the North, while Clegg is -60 or so.

    h ttp://

  12. ALLAN

    I’s fair enough. AW allows a pretty broad discussion for those of us who don’t come here for lessons in detailed statistical methodology.

  13. @Allan Christie

    In addition to your 2 hold predictions I don’t think anyone is expecting Charlie Kennedy to lose either. Even on 4% of the vote he’ll hold on along with Carmichael.

    Caithness is the one I’m really unsure about. My gut feel is a hold for John Thurso but they are very idiosyncratic voters up there. The Lib Dem vote is actually an anti Toffee nosed laird vote. Ironic when in Thurso they are voting for a Laird.

  14. Catmanjeff

    Intriguing figures. Might be worth looking at them over a longer time period.

    We’ve discussed “data mining” on here before, but I think it’s time for the miners to get their pit boots on and give us more information on more sophisticated correlations with VI.

    I was intrigued by Paul Bristol’s description of Bristol NW? as “mainly made up of Working Class Labour voters and Middle Class Tory voters”, and I wonder in how many parts of the country that remains true.

    For example, I’d like to see the VI figures for AB voters working in the public sector, as opposed to those in the private sector, and whether any differences match voting patterns in seats like East Renfrewshire, compared with seats with similar demographics elsewhere that elect Tories.

  15. @Norbold

    Did you see the little poll at the end of the article you linked? “Who will you vote for?” Top left – Farage…top right – Miliband…bottom right – Cameron…bottom-right – Clegg (with ‘Other’ over hid head).

    Proof once more that Farage is getting too much emphasis from the media?

  16. @ Northumbrianscot

    I found that a very illuminating set of arguments and projections. You calculate that there are ‘not enough’ LibDem votes in the pool to let them do as well as some expect and your final predictions take this into account.

    As it happens I have been grappling with similar problems at the UK level and I am still wondering how to handle it. In the UK-wide polls we know that the LD vote share has gone down from about 23% in 2010 to a VI of about 8% now. It is arithmetic ally impossible for this 15 point VI drop to have come from each and every constituency on an equal (= uniform) basis. Many constituencies had <15% share in the first place and in these cases there are 'not enough' 2010 LD to contribute their cut of the drop. It follows that a higher than average portion of the drop must have come from the stronger LD constituencies. At this point I did a series of regression analyses on Ashcroft constituency polls and confirmed that across constituencies the drop had a high proportional component: the LibDems were indeed losing more VI in their stronger seats.

    This introduced a new problem. Applying this formula across the country 'predicted' that the LibDems would lose every single one of their seats in May. As this seems implausible it suggests there is something wrong with fixed-proportional-drop as well. The next set of Ashcroft polls gave a clue. In LD seats using the conventional Voting Intention question the drop was indeed proportional (and very well predicted by my formula). But the figures Ashcroft headlined and everyone used were those for the *constituency-prompted* VI ("thinking specifically of your own constituency…"). These swung the VIs in these seats sharply back in favour of the LDs yielding several HOLD predictions. This looks like a kind of incumbency effect. The LD voters say they will go elsewhere until reminded of the circumstances in their own constituency.

    So – to return to your post – part of the reason that there are 'not enough' LD votes to save their Scottish seats may be that the country-wide polls are misclassifying these incumbent-led voters. In wider polls they go AWOL but rally round the sitting MP when given a suitable reminder.

    Having said all that, I wouldn't be surprised if a lot of the models overestimate LibDem resilience in Scotland, and I have a shrewd suspicion that the forthcoming Ashcroft Scottish constituency polls may prove to be a body-blow for the party. The reason for this is that I suspect that the Scottish context Ashcroft's Constituency VI question will prompt thoughts not of the sitting LD MP but of the resurgent SNP presence. If this happens then the incumbency bonus will be blown away, leaving the LibDems with much reduced support in the polls.

    On this basis my precision for the Ashcroft polls that the Scottish LibDems will be reduced to holding their 2-3 strongest seats.

    This doesn't mean that this is what will happen in the election itself. I don't think we yet know where the SNP VI trend is heading and everything may look quite different again in the spring.

  17. @CMJ

    Interesting weekly churn plots as ever. A hint of return to nurse for the Tories, with the LibDems still on walkabout trying to find a suitable resting place. Let’s try the Greens this week…

  18. Natalie Bennett will be toast in leader debates if she can’t improve on her performance on Sunday politics, she was evasive, and left many questions unanswered. Caroline Lucas comes across more of a peoples person, Natalie came across like the rest of them in Westminster.
    I wonder if the public see some of their pro- European policies and some of their more “radical” policies, like allowing you to be members of a terrorist organisation, and getting rid of our military sector and importing arms, whist shrinking the military to virtually nothing, that their VI may suffer, (a bit like happened to the BNP) when they came under scrutiny.

  19. @ Statgeek

    Useful charts again. Looks like a hint of a Ukip rise in the 15 poll graph for Scotland. Must look at that.

    The SNP 30-poll plot looks familiar. From August onwards I have tried to use a quadratic fit to describe this inverted-U curve. The fit is quite good so far but – if extrapolated – it predicts a catastrophic VI drop in the near future. On current evidence this seems rather implausible and so the kind poster who described the equation as a ‘spoof’ may be uncomfortably close to the mark.

  20. I notice that there is widespread support for the so called Mansion Tax which is not surprising since taxes which do not adversely affect the great majority are always popular. What is disturbing (apart from the class hatred shown) is that it is not at all clear how this will work and whether (after costs) it will actually raise much money. Ed Balls blustered his way through an explanation which sounded like he was making it up as he went along. At least there was slight recognition that there could be many such owners who do not have the spare income to pay the tax and the answer to that was that it would be added to their inheritance tax when they die. What a cheek! There will be endless arguments about the value of properties and no explanation as to who will pay for the valuation. If the value is just above £2mn what happens if the tax lowers the value of the property? It promises to be a typical Labour (or Lib-Dem) shambles. I should say that my property is not in the affected bracket! The Tory proposal to skew the Stamp Duty (whilst adversely affecting South-East owners like me) is much fairer since at least it is levied on real values and will produce a certain tax take.

1 4 5 6