Ten weeks to go

Here are this week’s polls.

YouGov/S Times (20/2) – CON 33%, LAB 34%, LDEM 8%, UKIP 13%, GRN 6%
Opinium/Observer (20/2) – CON 35%, LAB 33%, LDEM 6%, UKIP 15%, GRN 7%
Populus (22/2) – CON 32%, LAB 32%, LDEM 9%, UKIP 15%, GRN 6%
Ashcroft (22/2) – CON 32%, LAB 36%, LDEM 7%, UKIP 11%, GRN 8%
Survation/Mirror(23/2) – CON 28%, LAB 34%, LDEM 10%, UKIP 19%, GRN 4%
ComRes/Mail (23/2) – CON 34%, LAB 32%, LDEM 8%, UKIP 13%, GRN 8%
YouGov/Sun (23/2) – CON 33%, LAB 33%, LDEM 8%, UKIP 13%, GRN 7%
YouGov/Sun (24/2) – CON 35%, LAB 33%, LDEM 6%, UKIP 14%, GRN 7%
YouGov/Sun (25/2) – CON 33%, LAB 33%, LDEM 8%, UKIP 15%, GRN 6%
YouGov/Sun (26/2) – CON 33%, LAB 34%, LDEM 8%, UKIP 13%, GRN 6%
Populus (27/2) – CON 31%, LAB 33%, LDEM 9%, UKIP 16%, GRN 6%

The voting intention polls are continuing to show the same stasis we’ve had for the whole of the year so far, Con and Lab almost neck and neck, Labour just a smidgin ahead. Of this week’s polls five showed Labour leads, three Tory leads, three with a draw. The UKPR polling average is wholly unchanged from last week, remaining on CON 32%(nc), LAB 33%(nc), LDEM 8%(nc), UKIP 15%(nc), GRN 6%(nc). Perhaps the most notable change among some very unnotable polls was a change in who commissioned them – ComRes had been the pollsters for the Independent since 2006, but this week switched their monthly telephone poll over to the Daily Mail (they will continue to carry out online polls for the Independent’s Sunday stablemate).

Scottish, London and Constituency polls

TNS put out a new Scottish poll this morning with topline figures for Westminster voting intention of CON 14%(-2), LAB 30%(-1), LDEM 3%(-1), SNP 46%(+5), UKIP 3%(+1) (tabs). The previous TNS poll had shown an SNP lead of only ten points, this TNS poll is far more similar to the Scottish figures being shown by other companies.

YouGov put out a new London poll earlier in the week for the Times with topline figures of CON 34%, LAB 42%, LDEM 8%, UKIP 9%, GRN 6%. This gives Labour an eight point lead in the London, but given they won the vote in London at the 2010 electon is actually a slightly smaller Con>Lab swing that in the country as a whole. I wrote more about the poll here.

Finally there was a new Survation poll of Thanet South for UKIP donor Alan Bown, showing Nigel Farage with an eleven point lead. This compares with the Lord Ashcroft poll of Thanet South last November that had, once corrected, shown Farage one point behind the Conservatives. It may be that UKIP have managed to open up a lead in Thanet South since November, but there were also substantial methodological differences between the two polls – the new Survation poll prompted using the candidates names, which may well have helped Nigel Farage as the most well known of the candidates. There were also differences in weighting – Lord Ashcroft weights by recalled vote and by social class, whereas Survation don’t; Survation weight by council wards within the constituency whereas Ashcroft doesn’t. Finally there were don’t knows – Survation exclude them, Ashcroft assumes some vote for the party they did last time. And of course, this is a poll commissioned by a party – that should make no difference to how the poll is done (apart from adding candidate names this is Survation’s regular methodology), but it brings with it publication bias: if parties commission polls and don’t like the results, they don’t publish them.

Week 8

  • Jack Straw and Malcolm Rifkind were caught in a newspaper sting on MPs taking second jobs. Rifkind stepped down, Ed Miliband promised a ban on second jobs. YouGov polling found only 26% thought that MPs having second jobs helped keep them in touch and was better than full time politicians, 60% thought they should concentrate on their main job and second jobs risked corruption. 54% would support a ban on MPs having second jobs.
  • Immigration figures came out showing net immigration way above David Cameron’s stated ambition to reduce it to “tens of thousands”. I suspect the Conservatives failure to meet the target has long been accepted by the public and priced into their opinion though – early last year the proportion of people thinking it was likely the government would hit their target had already fallen to just 9%. Still, coverage of immigration will likely keep UKIP’s strongest issue high on the agenda.
  • Labour announced their policy on tuition fees. On the principle of who should pay for higher education the public are actually quite evenly split – 43% think it should be paid from general taxation, 42% that students should pay it through tuition fees or a graduate tax. For a reduction in the level of tuition fees though I expect Labour will get the thumbs up – in December YouGov found people were in favour of a reduction in tuition fees by 54% to 21%, even if it meant less funding for universities
  • And the debate debate struggled onwards. At the weekend the papers quietly suggested that the debates may now be dead, on Monday the broadcasters announced the order of the debates (the two big ones first, followed by the Cameron-v-Miliband head to head). For the moment though, it seems to have gone quiet.


The latest forecasts from Election Forecast, May 2015 and Elections Etc are below, along with the Guardian’s new election projection. As usual, everyone is projecting an extremely hung Parliament, with the two main parties close together in seat numbers.

Elections Etc – Hung Parliament, CON 279(-2), LAB 283(+1), LD 23(nc), SNP 40(nc), UKIP 3(nc)
Election Forecast – Hung Parliament, CON 285(+3), LAB 276(-4), LD 27(+2), SNP 39(-1), UKIP 1(-1)
May 2015 – Hung Parliament, CON 270(+4), LAB 271(-4), LD 26(nc), SNP 56(nc), UKIP 4(nc)
Guardian – Hung Parliament, CON 275, LAB 271, LD 27, SNP 51, UKIP 4

375 Responses to “Ten weeks to go”

1 2 3 4 5 8
  1. RAF

    Well, we’ve had a few 3-way marginals for a good few years now, and lots of folk vote tactically.

    Sometimes that has knocked one of the parties out of the race in succeeding elections – sometimes not, as tactical votes can cancel each other out.

    So folk decide their vote in the same way they decide most things – their “feel” for the way the constituency is headed, views of family/friends, how much do they dislike Party X compared to Party Y etc.


    ” In previous polling, however, YouGov has found that British people are divided over whether the cost of university education should mainly be paid for through general taxation (43%) or by students themselves (42%).”


  3. Northumbrianscot

    Together with Margaret Curran.

  4. The local paper has a poll showing Labour on 33, SNP on 31 and Tories on 27.
    I have always thought that a lot of the Labour vote comes from SNP voters voting tactically. Hopefully this will change at this election.
    My comment at 10.12am has a link to the article and may be in moderation for the rest of the day. :)

  5. Well we do have one poll for East Ren so far.





  6. @Anthony

    “Frequently asked Questions”.

    Now, if I didn’t know that you were a professional pollster, I’d say that thou doth protest too much! You’re right of course, and the nation is full of poll-deniers, especially if the results don’t serve their partisan preferences. I wouldn’t worry too much about it though, because it’s all the fun of the fair and a great addition to the gaiety of the nation.

    The one you could have added to your list of well worn myths, however, was our old favourite, privately conducted polling. For a party lagging behind in the polls, this is always, mysteriously and conveniently, much better for them than the public opinion polls.

    Strange that, don’t you think?

    Mind you, probably best to ignore the polls at the moment. It’ll all be all right on the night, as they say, with swingback occurring some time late in the afternoon of Thursday, May 7th.

    :-) :-)

  7. Allan Christie

    I wonder where they got those “polls” from? Maybe an old Election Forecast prediction? (currently L39 C28 S26).

    However, it will be interesting to see how much these individual seat projections are going to be used as “objective” justifications for “Only Party X can beat Party Y here” campaigning.

  8. @RAF

    One consequence of the ‘poor expectation management’ of the SNP, is the all the local parties think they can win the seat – so no tactical voting not even in East Ren.

  9. I wonder where posts go to when one’s browser decides to cut out mid scribble? No matter, @DAVE , In answer to your earlier question I would say about 35. Perhaps not a representative sample of the electorate though.

  10. @Old Nat

    You have to ask why he hasn’t engineered a by-election?

    1 He doesn’t have enough authority inside Scottish Labour!
    2. Not sure he would win a Holyrood by-election ?
    3. Not sure Labour can keep East Ren without his personal vote ?
    4. Doesn’t actually want to give up Westminster?

    Maybe Amber knows the answer but she is unlikely to tell us

  11. rmj1

    “I wonder where posts go to when one’s browser decides to cut out mid scribble?”

    What a fascinating philosophical question!

    I suspect that they enter a primordial soup, seeking to chemically combine with other partial posts.

    When they are sufficiently incoherent (but stable) they then appear on “Comment is Free” or football blogs.

  12. @Dave

    I live in Scotland and I don’t know anyone that votes Tory but for most of my life I’ve lived under a Tory government – How is that even possible?

  13. I didn’t suggest that Murphy intended to engineer a by-election – just that he hadn’t.

    The constitutional requirement for the leader of the LiS branch to be an elected member of a Parliament was always going to be predominant, but press speculation always provides the oxygen of publicity for a politician, so is “a good thing”.

  14. Thinking about the EU question, there’s a range of opinion from UKIP, via the Conservatives to LD’s on commitment to staying in. Oftentimes the Conservatives have pictured themselves somewhere between the two of getting out or staying in with the current settlement; thus the promises of renegotiation. It’s just that I think the time for such fence-sitting is steadily coming to an end. It may well be on us now.

    About Nick Clegg’s talk of not joining up with the Conservatives – it may just as well be that he does not wish to join a government that will still be a minority: presumably he’d rather have influence from outside the government. He may still offer C&S if that’s possible.

  15. Polldrums on the surface but with some interesting currents a little beneath the surface.

    It is quite remarkable how polldrummy (?) things are at the moment. “10 weeks to go” the thread title tells us. Look back the same distance and you see that then – as now – Anthony and UKPR contributors were talking about a slim 1-point Labour lead. Then – as now – the battleground seats were those in which the Tories had an 8% lead in 2010.

    The sheer predictability of most – though not all – VI movements can be seen by comparing Anthony’s current Polling Averages with the linear regression based projections I made in my original post on trends at 6.15 pm on Nov 24th.

    In that comment I gave the following equations to “predict” various party VIs on Day D (where D was zero on Nov 24 and is 99 today):

    Conservatives: Predicted VI on Day D = 31.97 – 0.00253 x D
    Labour: Predicted VI = 33.78 – 0.0138 x D
    LibDems: Predicted VI = 7.37 – 0.0079 x D
    Ukip: Predicted VI = 15.86 + 0.0113 x D

    The corresponding formula for the Greens was posted a couple of days later at @Ben Foley’s request:

    Greens: Predicted VI = 5.41 + 0.0098*D

    If we plug in the relevant D-values for the 21 polls in the current UKPR Polling Average list (range 79-98) you get the following comparisons:

    Party -Trend-based prediction: UKPR ave (discrepancy, + scores mean trend > actual average)

    CON – 31.7: 32.2 (+0.5%)
    LAB – 32.5: 33.3 (+0.8%)
    LD – 6.7: 8.2 (+1.5%)
    UKIP – 16.9: 14.5 (-2.4%)
    Greens – 6.3: 6.4 (+0.1%)

    So, the Trend-based projections from over three months ago have proved to be remarkably accurate for three of the parties. Also, Labour’s margin at this point was predicted to be just 0.8% and the actual figure is a statistically indistinguishable 1.1%. In these cases it is almost as if no further polling data has been required over the past three month. The figures from before that told us all we needed to know about where we would be now.

    If this pattern continues as it has in the past then the seat tallies in the Election (posted earlier in this thread) will be rather more Labour-favouring than most of the public models indicate.

    Perhaps more interesting than the accurate aspects of the trend projections, though, is the pattern revealed in the LD and Ukip departures from trend.

    Told in advance that LD VIs would rally a bit in early 2015, I suspect that most UKPR contributors would have expected to see a corresponding drop in Labour’s support: if votes stick to (or return to) LD support, Labour would be expected to lose out a bit relative to the Tories. Equally, a topping-out (or even decline) in Ukip support has always been expected to advantage the Tories.

    However, while both of these minor-party changes have happened since Jan 1, there is no sign at all that the Labour margin has been pulled away from trend. Whatever complex patterns of churn have taken place to produce these headline changes, they don’t seem to be having the overall effects that contributors like @Pressman and others were predicting.

    If churn expectations can’t be trusted, how can we foretell the future?

  16. And as for Swingback, still no sightings.

    As far as I am aware, the only advocate who sticks around here is @Number Cruncher. (@Robin Hood runs back into the woods as soon as he’s made his intermittent proclamations).

    A couple of months ago @NC was still observing that swingback was on track (see Charts 9 and 10 on his post).

    His projections indicated that Labour’s margins could be expected to drop by a full 3% between 5 months and 3 months out from the election.

    Reality: No reliable margin reduction at all.

    Verdict: Still waiting for Swingbot (with apologies to the contributor who first cited Beckett).

  17. @Anthony Wells – apparently, I’m on the naughty step – I’m unaware I said anything out of line but if I did, mea culpa, mea culpa, mea maxima culpa: sorry, John

  18. John Murphy

    It’s that damn confusion of you with Jim! :-)

  19. As regards Jim Murphy’s long term plans to get into the Scottish Parliament I still think the idea of him going head to head with Nicola Sturgeon in Glasgow Southside has some political merit.

    The 2011 result was 54% SNP, 35% Lab, 8% Conservative, 3% Lib Dem.

    So it only takes a 10% SNP to Labour swing to win it, which is the sort of numbers nationally Scottish Labour would be needing to have a chance of minority government.

    Less than 10% swing if Jim can squeeze some of the Conservative voters to unite against the SNP.

    It would be a brave tactic, especially if he didn’t stand on the list and made it an all or nothing choice “Whoever wins has the mandate to be First Minister”.

    If he lost he could resign as Scottish Leader and go back to his Westminster career.

    If he won he’d have undermined Sturgeon’s credentials even if she was returned on the list and SNP had more MSPs than Labour.

    High risk, high reward strategy but might suit him if he fancies being First Minister but not so keen on leading the opposition at Holyrood for 4 years.

  20. “only takes a 10% SNP to Labour swing to win it”

    We’re in a strange world when a 10% swing is ‘only’!

  21. Unicorn – excellent work as always and I understood it all this time (Or at least I think I did!).

    Just one comment I query, you said:

    ” Equally, a topping-out (or even decline) in Ukip support has always been expected to advantage the Tories.”

    I think many of us on here who have expressed a view thought the last in first out principle would apply.

    It is undoubtedly the case that the UKIP increase of say 13% has been at the expense of the Cons most so if they fell all the way back to 2% *which wont happen) the Tories woulld gain 4-6% over Labour of GB VI.

    The evidence seems to suggest, however, that whilst the tranche of increase to around 10% came early in the parliament was disproportionatley from teh Tories perhaps 4% more than from Labour; and, that the later movememt was much more evenly split.

    The suggestion from me and others is that the longer the VI has been in place the harder it is to shift back so that if the UKIP decline to around 10% GB wide the Tories will benefit little over Labour.
    The hypothesis contends that the UKIP VI needs to fall below 10% for the Cons to see a significant net gain over Labour in vote share terms.

    Of course, my caveat is that in marginals the UKIP will get squeezed more which will probably break for the cons but that the greater squeeze on the Greens in such seats will nullify this.

  22. Well I know it’s a lot but if there isn’t a 10% swing then he won’t be First Minister anyway.

    I personally don’t think it would happen but could see why it might appeal to Jim Murphy.

    A not dissimilar tactic was used by Alex Salmond in Gordon in 2007.

  23. @ Northumbrianscot,

    Surely there isn’t a hope in hell of Labour taking a currently SNP-held Glasgow seat. They’re going to be lucky to hold Glasgow seats they held with 50% majorities in 2011, much less pick one up from a more popular party leader.

    Although he was willing to take on the leadership of Scottish Labour I don’t think Murphy is quite that big a masochist.

    @ Unicron,

    Huh. I’d thought that the Lib Dem recovery was coming at the expense of the Greens, but apparently not. Is it all from Ukip? (Well, I guess we’ll know when I finish February’s churn analysis.)

    Good Afternoon to you.
    In addition to the thrice repeated culpa; we had to bring in the angels and saints of various degrees in days of old, and then to do penance and gain some sort of indulgence.

  25. @Spearmint
    “Well, I guess we’ll know when I finish February’s churn analysis”.


    Where’s all the “interesting”, “pollercoaser”, ” humdinger” hyperbole?

    I’m afraid Spearmint, that as brilliant as you are, you just do not have the required marketing skills to be a pollster ;)

  26. Also, unless Jim Murphy has a ridiculously inflated view of his own competence or the SNP’s incompetence- admittedly, a possibility we can’t rule out- I think we have to assume that he intends to lead a Holyrood opposition for at least five years. It’s a bit like that Michael Foot line about how no one who joined Labour under his leadership can be dismissed as an opportunist. No sane person could have looked at the position of Scottish Labour at the time when Murphy decided to stand for the leadership and thought “I’m definitely going to be First Minister in 2016”. The job vacancy was for opposition leader.

    So the idea that he’d risk his career on an all-or-nothing gambit like this because he can’t stomach the thought of opposition strikes me as pretty fanciful.

  27. @ RAF,


    Sad but true.

  28. @JimJam

    It would certainly be interesting to see whether there is support for your rather sophisticated ‘layered churn’ hypothesis (that churn patterns from Ukip will follow one template down to 10% and then a different structure below that).

    (In passing, I note that @CMJ’s YouGov-based churn plot didn’t support your hypothesis for the Jan/Feb changes. Not a hypothesis-slayer: just an observation…)

    For the record, in my tweaks to the trend-based seat projections, I have assumed that the Ukip VI will now continue to drop at the new 2015 rate of 1.9% per month. That would take it down to your suggested 10% level and not much below it. So, on your analysis it would seem that @Pressman shifts will not happen in this election.

    In my own calculations – following the Ashcroft 2nd preference poll – I move VI from Ukip to Tories/Labour in a ratio 2-to-1.

    I sense that it will be these very subtle and complex details that may determine the outcome of the election.

    Finally, for those on this thread who feel I might have forgotten Scotland, my current calculations assume that Scotland is done and dusted. I have searched – without success – for post-referendum trends, and as far as I can see everything was somehow settled by November, and we are going to go into the election essentially with the VIs posted diligently by @OldNat. That’s the assumption that goes into my GE projections.

  29. @Unicorn

    I’m something of a swingback advocate, but I think if and when it happens, it’ll be quite a modest change to current trend Labour 1% lead – I think at most a 2% Conservative lead. Just about enough to get them largest party status, but with a real struggle to find enough allies to form a government. Very much not enough to win them an OM, slim or not.

    If we are still in a situation like this say in 4 or 6 weeks time I’ll be one of those scratching my head and “wondering where’s the swingback?” – I’m thinking I’ll be doing that soon enough.

    It’s not a given that large trend polling changes occur in the period leading up to an election: out of the last 30 years worth of GE’s, several times it was not really worth bothering with a campaign: they might as well had the poll 2 months early for all the difference the extra time made.

    We shall soon find out whether all this fiddling around with predictions gubbins is worth the bother.

  30. Unicorn

    As you know, I don’t claim that my crossbreak means are “right” – just that if there was any change developing, I’d expect to see some sign of it in them.

    For the record, the weekly YG mean scores for SNP/Lab in 2015 are 41/27, 42/27, 41/28, 44/27, 41/25, 43/27, 42/26 – and I don’t envisage any change tomorrow.

    The phrase “settled will of the Scottish people” comes to mind – though not as the late John Smith imagined it.

  31. Unicorn – the last in first out part of my hypothesis (not just mine) is the debatable element, I acknowledge.
    I trust you will acknowledge alos, though, that your initial comment suggesting it ‘has always been expected’ that a declinling UKIP VI would benefit the Cons is based on a broadly linear returning home from the UKIP which is equally uncertain; and, not a consensus opinion on this board alkthough one held by some contributers I accept.

    That the UKIP VI broadly grew at the expense of the Tories significantly up to AROUND 10% is I believe supported by the evidence from the polls for the first 2 years or so of the parliament
    (some may have gone Con-Lab for a time due to omni-shambles for example before settling on UKIP but that would never have stayed Lab through to 2015 anyhow).
    Similarly the last 5-6% of the UKIP VI growth being more evenly spread is supported I recall by polling evidence but I am not going to trawl through to find it.

    I missed CMJs churn analysis (was away a week in Feb), it is able to detect specific drfits from one party to another or just record net outcomes of various movements?

  32. @ John Nolan

    I am a Canadian political scientist, John, observing these elections from afar. Can you tell me by rough estimates (%) how those people voted in 2010 and when they started switching to voting UKIP, which election?

    Many of the pollsters are weighting to 2010 and I have repeatedly said on this listserv that it is a staitically erroneous thing to do, because it fails to recognize that voting patterns have changed.

    In 2010 1: 33 voted UKIP, but in the 2014 European election it was closer to 1:3.

  33. Keith P

    FWIW, I am with you and the SNP surge is what makes the cons largest party the likely reuslt imo with a 2-3% lead which most likely would not have been enough otherwise.

  34. The local paper has a poll showing Labour on 33, SNP on 31 and Tories on 27

    It would have been helpful to know which poll the local paper is referring to. Does anybody know the provenance of this polling?

  35. @ UNICORN

    No disrespect intended but is it not still possible that a statistical model will work as predicted, but at the end of the day still be wrong in terms of the reality of outcome.

    I agree with your statistical model as published at 12.04 PM, but I still question the veracity of the model itself. Not yours per se, but the models presented by the pollsters.

    I remain unconvinced that weighting for 2010 is an accurate way to determine voter intention in the 2015 UK election, especially when the number of people voting UKIP, for example, has gone from 1:33 in 2010 to roughly 1:3 in the 2014 European election.

    Note the difference in South Thanet between Ashcroft and Survation, when the 2010 weighting is turned off.

    What happens to the model if you turn off the 2010 weighting, where would we be now if we just asked how are you going to vote and discarded all “don’t knows”. “refused to say” and “will not vote”?

    What’s the picture among those who are willing to tell us directly what they intend to do on May 7th.

    Are the pollsters putting on so many statistical filters that they cannot see the real trend, a bit like wearing a set of rose tinted glasses.

    Might it not be better to say here is what 75% of those polled are thinking, but we really do not have a good ideas about how the other 25% might vote.

    Is that not more intellectually honest, for the pollsters, than to posture and second guess what might be happening based on an election before UKIP started electing local government councillors and the largest pluarlity of MEPs?

    Both the French and Russian aristocracy sat around pretending that a revolution was not coming, though the cabinet minutes for both governments clearly indicated both knew it was.

    Human beings hate change as we are creatures of social habit, so I keep wondering about whether pollsters are engaging in statistical denial – hence the occasional tweak that this or that pollster has “tweaked” their model.

    No one wants to be the bearer of bad news, like telling LD they are going to be wiped out or Labour that their support is going backwards or the Conservatives that their ‘safe” seats are really no longer safe at all.

  36. “If this pattern continues as it has in the past then the seat tallies in the Election (posted earlier in this thread) will be rather more Labour-favouring than most of the public models indicate.”

    Swingback theory, for want of a better phrase, has been constantly under pressure since Fisher’s first projection in September or October 2013, when he suggested that the Tories would get 40% and a majority of about 25, I seem to recall, (something like 337 seats)…many of us questioned this at the time, and were highly sceptical. As of yesterday his model is suggesting that the tories would get 279 seats, while labour would get 283.

    His model today still presupposes a significant amount of swingback, as it implies the tories on 33.6% and labour on 31.%. I still believe that labour will be much closer to the tories than fisher is suggesting. His model has consistently over-estimated the level of swingback which will occur and I don’t see why it would suddenly be correct today on 28th February.

    The labour VI is quite solid, as are those of the other parties. I think 31.5% is too low an estimate for them, given where the polls are currently. The penny is slowly beginning to drop that the Conservatives are unlikely to form the major part of a sustainable government post May.

  37. @Andy

    I think the general thought here is that EU elections are very different to General Elections

    If you look at the 2009 EU elections, the results were
    Cons 28%
    UKIP 17%
    Lab 16%
    LD 14%
    Green 8%
    BNP 6%

    Only a year later in the general election
    UKIP 3%
    BNP 2%
    Green 1%

    So the small parties disappeared.

    Different levels of turnout
    People don’t vote for small parties in general elections as they think they are wasted votes
    People are voting for different things

    Yes, UKIP and the Greens seem like they may be more resilient this time, they are holding up better in the opinion polls, but we can see even at the time of the EU elections, people intended to vote differently in the EU elections vs the general election


    Press view data table to see the figures which are easier to follow. Of the people who voted UKIP in the EU elections, back in May

    23% planned to vote conservative
    10% planned to vote labour
    52% planned to still vote UKIP

    in the General election.

    While I agree re-allocating don’t knows based on 2010 may be wrong, re-allocating based on EU results would also be wrong for the above reasons.

    Hopefully they start making their minds up soon so we don’t have to guess what all these don’t knows are going to end up doing.

  38. @ Richard

    I think what disturbs me most is that the pollsters are using voting intention from only one election to weight.

    I agree that just because someone voted UKIP in the Euros in 2014 does not mean that they will vote UKIP again.

    But if somone has been consistently voting UKIP then I think there is a higher probability they might do it agin in the GE.

    What I am after statistically speaking is a better probability analysis, including understanding the crossover between UKIP, BNP and ED. I think UKIP has a much bigger core than is agreed to on this listserve.

    I think that a swath of UKIP and Green voters are getting ready to switch from voting that way in European and local government elections to voting that way in the 2015 GE.

    The same pattern has emerged for SNP in Scotland, where an increasing number of voters having voted SNP in the Scottish Parliament elections and local government elections and are now signalling that they will do so in the GE.

    I find it worth observing that UKIP voters are more determined to vote than Labour of Conservative and that their firmness to vote UKIP is as strong as Labour and Conservative voters.

    In contrast Green voters while stronger in their intention to vote, still indicate roughly half could go elsewhere. Among LD that figure rises to over 60% and they are far more disheartened.

    I watched the emergence of a socially-conservative and libertarian economic party in Canada between 1988 and 1993: Reform.

    UKIP is not as strong as Reform were, but I still expect to see some very tight races on May 8th in the Southeast and Midlands.

  39. @Andy

    Personally, I think a better way of allocating don’t knows would be to look at where the other 2010 voters have gone since then, and weight them in the same ratio.

    So 2010 Conservative unknowns should go 70-80% conservative, 20% UKIP
    2010 Lib Dems should go 30%ish LD, 30%ish Labour, 10%ish Cons, rest to Greens, etc…

    That then still looks at how they voted in 2010, but also looks at the new political landscape.

  40. Test

  41. Interesting political strategy by UKIP:


    Presumably they have decided Scotland is a lost cause for 2015 but maybe they should be thinking about 2016!?

  42. @RMJ1 Thanks.

    @Dave I live in Scotland and I don’t know anyone that votes Tory but for most of my life I’ve lived under a Tory government – How is that even possible?

    Why have you asked me?

    However I’m fairly sure that there are many more people in the north of England who could say the same thing.

    It is possible because since 1945 about half the people who voted, voted Tory and UK has had governments alternately Conservative and Labour which reflect this. Of the 70 years since 1945 31 have had Labour governments. “Most of your life” is true, but really means “rather more than half” doesn’t it?

    If you really do not know anyone who votes Tory, perhaps you should widen your circle a bit. Tories do have a few good ideas.

  43. On another note, I’ve heard rumours of a constituency poll of Bristol West putting the Green Party 2 points ahead.

    My question would be that if it exists, why haven’t we seen it? Unless it comes out soon… unless it wasn’t conducted by the Greens and is being sat on.

  44. Not conducted, paid for. Of course.

  45. @KeithP

    I’m something of a swingback advocate…

    Yes – I remember you ‘lumpy swingback’ thesis (which someone suggested might be a reference to an East End thug).

    Although you say you might start wondering where it is, it’s unpredictability makes it difficult for others to call you to account.

    That said, if I were to present myself as a trend advocate, the recent Ukip and LD shifts would have to make me think twice about that. The EF team would regard these developments as being in line with regression-to-mean and because if this might well argue that swingback is happening already.

    However, I suspect the clincher for most people would be the appearance of a steady and substantial rise in Tory VI, perhaps even more marked than the change you are expecting.

  46. ALEC


    My that’s a purley little post.

  47. @JimJam

    I miissed @CMJ’s.churn analysis [..] able to detect specific drfits from one party to another or just record net outcomes of various movements?

    Yes – because he tracks changes using voter IDs. You can find his most recent histograms here. Inconsistent with my comment above, they suggest that the bulk of recent Ukip loss has been returning Tories.

    @Spearmint is apparently on the point of posting another churn analysis and I, for one, am looking forward to poring over those details too.


    The local paper has a poll showing Labour on 33, SNP on 31 and Tories on 27
    It would have been helpful to know which poll the local paper is
    referring to. Does anybody know the provenance of this polling?

    I’ve sent an email off to the Barrheed News asking them to authenticate the poll but if I’m being honest it’s probably about what I would expect at this stage.

  49. @ Richard

    Yes, I think I could live with a formula like that, as I have seen swinback or voter oscillation as an elections nears – but believe both UKIP and green have a core that is totally different from 2010, and the same is true of the LD as you correctly poinbt out in your last post.

1 2 3 4 5 8