Eight weeks to go

Here are this week’s polls

YouGov/S Times (6/3) – CON 34%, LAB 33%, LDEM 8%, UKIP 15%, GRN 5%
Opinium/Observer (6/3) – CON 34%, LAB 34%, LDEM 8%, UKIP 14%, GRN 7%
Populus (8/3) – CON 32%, LAB 33%, LDEM 9%, UKIP 15%, GRN 6%
Ashcroft (8/3) – CON 34%, LAB 30%, LDEM 5%, UKIP 15%, GRN 8%
YouGov/Sun (9/3) – CON 35%, LAB 31%, LDEM 8%, UKIP 14%, GRN 6%
YouGov/Sun (10/3) – CON 33%, LAB 31%, LDEM 8%, UKIP 15%, GRN 6%
YouGov/Sun (11/3) – CON 34%, LAB 35%, LDEM 7%, UKIP 14%, GRN 5%
Ipsos MORI/Standard (11/3) – CON 33%, LAB 34%, LDEM 8%, UKIP 13%, GRN 6%
YouGov/Sun (12/3) – CON 33%, LAB 32%, LDEM 7%, UKIP 16%, GRN 6%
Populus (12/3) – CON 29%, LAB 32%, LDEM 8%, UKIP 18%, GRN 6%

After two months of very little movement there is finally some sign of a shift. YouGov’s daily poll seems to have flipped over into showing a Tory lead – four of their five polls this week showed a Conservative lead. Populus is still showing a Labour lead, as did the monthly Ipsos MORI poll this week, but it’s the underlying trends that count and, for now at least, the UKPR polling average puts the Tories into a tiny lead – CON 33%(nc), LAB 32%(-2), LDEM 7%(nc), UKIP 15%(nc), GRN 6%(nc).

Scottish and Welsh polls

YouGov also produced Scottish and Welsh polls this week. The Welsh polling is here, and had topline Westminster figures of CON 25%, LAB 39%, LDEM 5%, Plaid 10%, UKIP 14%.

The Scottish polling is here and has topline figures of CON 18%, LAB 27%, LDEM 4%, SNP 46%, UKIP 2%. The SNP lead is down two points since the previous YouGov Scottish poll at the start of February, but remains at a formidable 19 points.

The Scottish poll also included some questions designed to tease out how effective Labour attempts to fight the SNP with a “vote SNP get Tory” sort of message. The short answer is not very. Essentially, for such an approach to work SNP voters would need to believe that returning SNP MPs really would make a Tory government more likely, would need to think a Conservative government was significantly worse than the alternative and would need to consider avoiding a Tory government as more important to them than the opportunity of returning lots of SNP MPs. All of these requirements are quite weakly represented amongst SNP voters – only 27% of SNP voters buy the argument that more SNP MPs will make a Conservative government more likely and while 38% of SNP voters think a Labour government at Westminster would be better for Scotland than a Tory one, only 15% think there’s “much” difference, and 49% think there’s little difference at all. Finally, even if SNP voters did think that returning SNP MPs would make a Tory government more likely, by 46% to 31% they’d rather have a Tory government and lots of SNP MPs than a Labour government and not many.

Week 10

  • The debate debate was still rolling at the start of the week, but has thankfully now faded out a bit. The rival bid to host the debates online by the Guardian, Telegraph & Youtube have said they they would host a five way debate before the end of March, so meeting David Cameron’s demands on timing. Otherwise there has been no further progress
  • There was debate earlier in the week about whether the government should commit to the NATO target of 2% spending on defence. YouGov asked about Defence spending last month and found 49% think we should spend more on defence, 16% less and 20% about the right amount. Defence spending is an interesting subject as it does at least divide opinion. Most spending issues are really one way streets – the vast majority of the public think spending on things like the NHS, policing and education is a good thing and would like more of it, it’s just a case of what governments can afford and how they fund it. Equally opinion polls asking what should be priority for cuts always find overseas aid top of the list. Defence is one of those few issues where there is both a significant chunk of people who think spending money on it is a bad thing that should be cut, and a significant chunk of people who think it is a good thing that should be protected. Welfare is another.
  • Nigel Farage got into a row over discrimination laws – appearing to say that UKIP would repeal discrimination laws, but later saying he meant only laws on discriminating on the grounds of nationality. When it comes to discrimination on grounds of nationality, I expect we will find most people agreeing with Nigel Farage – there has not been any polling on what he said yet, but looking back YouGov found in 2011 that 51% of people thought companies should prioritise British workers, even if there are better qualified foreign workers. What may be less good for UKIP is if the political row around Farage’s statement damages the party’s already somewhat shakey image on racism – last month ComRes found 44% of people agreed with the statement that UKIP were racist, up from 32% last year.
  • By the end of the week the main political issue appeared to be how many kitchens Ed Miliband has. I think it’s fair to say that’s the sign it hasn’t really been a huge news week. Onwards to…
  • Budget week. The budget is on Wednesday, one of the few big set pieces we’ve got before the election that we can reasonably assume most people really will notice, and which does have the possiblity to actually shift votes. As I say most years, people are often too ready to assume that the budget is an opportunity for the Chancellor, when often it’s a bullet to be dodged – there are plenty of instances of budgets damaging a government and not that many of them providing a real boost. Nevertheless, the timing means this is likely to be an extremely political budget and we shall see if it has any effect

Projections

The latest forecasts from Election Forecast, May 2015, Elections Etc and the Guardian are below (the Polling Observatory team’s projection should finally make it’s proper debut this coming week). All predict a hung Parliament, and all continue to predict the Conservatives winning the most seats.

Elections Etc – Hung Parliament, CON 285(-1), LAB 279(+1), LD 22(nc), SNP 40(nc), UKIP 3(nc)
Election Forecast – Hung Parliament, CON 288(+2), LAB 271(-9), LD 26(+2), SNP 42(+4), UKIP 1(nc)
May 2015 – Hung Parliament, CON 281(+5), LAB 263(-8), LD 24(+1), SNP 55(nc), UKIP 4(+1)
Guardian – Hung Parliament, CON 279(+4), LAB 265(-6), LD 27(+1), SNP 53(+1), UKIP 4(nc)


273 Responses to “Eight weeks to go”

1 2 3 6
  1. @Amber

    see last thread for an answer

  2. Surely it is no coincidence that the Budget is on Wednesday and on Friday we will be plunged into darkness, as the sun is blotted out from the sky?

  3. Can GO move VI with this budget?

  4. @Colin

    Vince Cable thinks GO has about £5bn to play with. What he does with it may decide the election.

  5. On the figures from all 4 projections, it is hard to see how a Government able to pass a budget could be formed.

    Can see a position where the result of a July Election would depend on parties abilities to form partnerships or agreements in the weeks after May 7th.

    Having said that, 8 weeks is a long time in politics. Giveaway budgets, scandals, amazing speeches, egg throwing farmers or kitchens could still affect things.

  6. Colin

    Possibly not but the press will achieve it for him anyway.

    This week has seen a tiny shift but I am not convinced we are seeing a permanent one, – yet.

  7. @AW

    I know it’s sensible to work with rounded figures. But if you calculate this week’s and last week’s weighted averages to a couple of decimal places Labour’s drop has been almost exactly one unit (rather than the 2 point drop you give in your summary).

    Also, I still can’t see why this week’s weighted average for Labour isn’t 32.5031, rounding up to 33% and ‘tied’ with the Tories’ figure.

  8. RAF

    I doubt it will change the election-but GO is capable of generating a small hike. It will fade though unless Cons have a positive Campaign in which they articulate reasons to vote Conservative.

    Howard

    How could anyone be convinced of any permanence in these Polls?
    It is all quite extraordinary.

  9. Budget week – my favourite week of all

  10. The Budget though has long ceased to be the special occasion that ir formerly was – when MPs such as Gerald Nabarro and Leo Abse would dress up for the occasion.

  11. OLD NAT
    Norse gods are at work and Celtic maidens beware. The Saxon is at his most dangerous. Woe woe and thrice times woe.

  12. Top ten most noticed news stories this week

    https://twitter.com/PopulusPolls/status/576434168520962048/photo/1

    Jeremy Clarkson 22%
    ISIS 6%
    TV debates 6%
    Helicopter crash 5%
    Girls travelling to Syria 4%
    Death of Becky 4%
    General election 3%

  13. Dear Old Nat — it would be a nice congruence and example of celestial power but we will not be having a total eclipse on Friday. Well unless you are in a boat off the coast of the Faroes.

  14. I have to say that I strongly suspect that some of the prediction models will turn out to have been a load of tosh. In my view they giving far too much weight to individual constituency polls and largely failing to take account of the margins of error that need to be applied to them.

  15. Abbreviated trend summary:

    This week the Tories have (surprise, surprise) shifted up above their long term trend, having been right on trend for most of this year. Labour and the Greens remain securely on trend. The LibDems are rising above their long-term trend for the ninth week in a row, suggesting a modest for of Swingback-like recovery. Ukip remain below trends as they have done for all but one week this year. Again, this might be described as a form of swingback.

    One event that did NOT occur this week is a Green/LibDem crossover. My original trend equations projected this as happening yesterday. But in the event the LDs are still 0.8 points ahead of the Greens (based on the figures in the current Polling Average list). What happened was that the LDs pulled out of their dive at the end of last year. @Spearmint and others have claimed that on YouGov figures there was a brief period of crossover. But I have never seen any evidence that the Green approach moved beyond MoE. As I see it the LibDems have lifted away a bit and will no longer be harried by the Greens (this time round).

  16. Good Evening All, after a good day in work, but home from a very frustrating chess match; snatching defeat from jaws of success.
    UNICORN. Very interesting take on the numbers thank you.

    COLIN.
    I agree that the numbers in the polls are strange. I think that Graham points out sometimes that when the Campaign officially starts the Government loses some VI, possibly due to greater media coverage of the parties in Opposition. 1983 and 1992 are exceptions I think. Therefore GO will make a big impact speech, possibly on National Insurance or Allowances for the low paid, IMO

  17. @ Unicorn

    Thank you. I take your point… but Ashcroft’s methodology is not ‘tried & tested’ because he’s the new kid on the polling block. Nor are constituency polls common enough to warrant the ‘tried & tested’ description; GE constituency polls which are available to the public are a relatively new development.

  18. Looking at these figures finally swinging to the cons, if EM costs labour the election then will labour MPs (the majority of whom didn’t vote for EM) feel entitled to exact just revenge on him on behalf of the nobly silent brother he stabbed in the back? Answers on a post card (or betting slip before the odds come down ;) )

  19. @Nostra

    If…

  20. @Unicorn

    Your findings sound remarkably similar to my earlier CUSUM analysis.

  21. @NOSTRA
    Looking at these figures finally swinging to the cons, if EM costs labour the election then will labour MPs (the majority of whom didn’t vote for EM) feel entitled to exact just revenge on him on behalf of the nobly silent brother he stabbed in the back? Answers on a post card (or betting slip before the odds come down ;) )
    ——————————–
    Why do you think David Miliband was stabbed in the back? Did he have some claim on the job of Labour leader? Ed offers a different perspective to his brother (who is tainted by Iraq) and the Electoral college voted for Ed’s viewpoint. He lost, he wasn’t stabbed.

  22. The projections given a Lab/Lib coalition with SNP C&S. I think that is a very likely outcome.

    There are two things that might change the result based on the polls Cons strength is the economy and Labour’s weakness Miliband.

    1. The budget, I know the bounce may be short lived but it could bounce the Cons into the campaign and set the Cons narrative. My prediction is:
    – Increase the minimum wage (combats Labour’s cost of living crises)
    – Tax breaks for middle earners either cut in basic rate of re-introduce the 10p tax rate

    2. Ed Miliband: He was very good on the free speech programme – if he can get exposure in that sort of setting we may get a version of Milimania. I know it seems unlikely but people have been judging Miliband on sound bites and the prism of the MSM, when they are in the mood to listen they might find him quite different as to how he has been portrayed

    So at the moment all to play for.

  23. Hookeslaw

    There is no beauty in your soul. :-(

  24. @Graham

    I have to say that I strongly suspect that some of the prediction models will turn out to have been a load of tosh. In my view they giving far too much weight to individual constituency polls and largely failing to take account of the margins of error that need to be applied to them.

    As you know, I am sceptical enough about the various models to keep picking away at them.

    But I think they have done more to handle MoE concerns than seems apparent on the surface. For EF at least, my understanding is that they don’t just plug in the Ashcroft result and take that as sacrosanct. For a start, the Ashcroft VIs are adjusted daily following their initial input. This causes the constituency balance to change over time in line with national and regional shifts. More importantly, the final seat tally for any given party is not a simple aggregate of all the seats in which they come out first. Rather, a probabilistic formula converts each constituency profile into fractional seats for each party and these are aggregated to give the final projection. This process is repeated thousands of time with each run giving slightly different counts, and the full range of outcomes are used to calculate the confidence limits of the final projection.

    What this means is that each seat is allowed to oscillate through the full range of its MoE, so that the indisputable inaccuracies of individual polls are averaged out in the eventual projections.

    So, to make it a bit more concrete if an Ashcroft poll gives a Labour VI of (say) 30 for some constituency, for some of the runs the figure will actually be entered as 27, 28 or 33 – indeed, any figure drawn from a populaton of 30 and standard deviation of 1.5. The same will happen for all other parties and all other seats. The end result is that effects of (undenied) polling variation are averaged out in the final projections.

    In short, variability is probably handled securely enough. The problem lies, however, in systematica biases – which is what I was exploring in that earlier post.

  25. Paul

    Just the way I remember it at the time but you just put forward quite a strong defence there for EM’s conduct and I’m in the ‘humblest student of politics’ gang I’m afraid so I may have read it wrong back then so fair comment.

  26. @CMJ

    Your findings sound remarkably similar to my earlier CUSUM analysis.

    Yep. My methods consistently lag behind yours. The pattern hasn’t escaped me. :-(

  27. I fail to understand why the seat forecast is 22 to 27 for the LibDems based on 7% to 8% but only 3 or 4 for UKIP based on 14% to 18%.

  28. @ couper2802

    There are two things that might change the result based on the polls Cons strength is the economy and Labour’s weakness Miliband.

    Some folk (mentioning no names) say that the Ed Miliband factor is already factored in.

    I’m not so sure.

    If it is, it is.

    If it isn’t, I think LAB will lose another 1% or so at the GE because of him.

    Remember the late anti-Kinnock swing in 1992? Some of us made a fair amount of money on CON winning that election and we did it by taking into account a late swing AGAINST Kinnock….

  29. @Simon

    Simple.

    The Lib Dems pile up votes in a small number of safe seats.

    UKIP draws votes broadly but thinly. Under FPTP that favours the Lib Dems.

  30. @ Simon

    I fail to understand why the seat forecast is 22 to 27 for the LibDems based on 7% to 8% but only 3 or 4 for UKIP based on 14% to 18%.

    No-one wins a seat with 14% or even 27%.
    Parties generally win seats by getting something like 45% or so in a given constituency.

    LD have about 25 seats where they will get that 45% or whatever it is. Across the rest of the country they will get more like 6%. Or less. So their National average is around 8%.

    UKIP, by contrast have a National average of around 15%. But only in perhaps 3 or 4 seats do they look like getting the 45% (approx) required to win a seat.

  31. @ David in France

    “Remember the late anti-Kinnock swing in 1992? Some of us made a fair amount of money on CON winning that election and we did it by taking into account a late swing AGAINST Kinnock.”

    I can’t see Ed holding a Sheffield rally the weekend before the election…unless Mr N is arranging something we don’t know about….

  32. David Cameron’s popularity was priced in to the 2010 GE; the Tories expected to win comfortably – until the last lap, when it turned out that his personal popularity wasn’t all it was cranked up to be.

  33. @David in France

    Exactly it could go either way. I don’t think people have really ‘got to know Miliband’ they haven’t been able to because of the poor start (beating his popular brother) then the negativity of the media. So Miliband is in Clegg’s 2010 position albeit with a bit more baggage and preconceptions

    It is perfectly possible that once people get to see more of Miliband Labour VI will fall OR people may begin to like him – that is why the Miliband effect during the campaign is an unknown. It is in my opinion the factor that will determine how Labour does in the campaign. Will people warm to Miliband? If they don’t then I agree 1992 beckons

  34. Personally I like Miliband a lot but then I liked Kinnock so I am not a great judge. However the online reaction to the Free Speech event was very positive, so I do have hopes for him.

  35. @Couper2802 @David in France

    Labour could go up or down?

    That prediction is really sticking your neck out ;-)

  36. David in France, interesting point about Ed there. But looking at his personal poll ratings I’d be very surprised if the Ed factor isn’t already factored in.

  37. Much speculation about the LDs siding with the Tories – but what sort of LDs will survive in England?

    Are they likely to be ones who see their personal survival in their seat as being enhanced or damaged by a further coalition with the Tories?

    I suspect personal survival may be a significant factor for many English LD MPs.

  38. Also isn’t one of the reasons people believe David Cameron doesn’t want to the debates is he doesn’t want Eds ratings to go up?

  39. @ Unicorn et al

    UKIP and the Green Party are not down if you take Scotland out of the equation and factor in England only and then only some regions of England like the Midlands and the South outside London.

    There is also a failure on the part of some to recognize that there are more than three parties in this election. In some seats the race will be between three and even four parties, so the threshold to win in some instances will be closer to 33% and in others potentially as low as 25%.

    A case in point is Brighton-Pavillion in 2010:

    Green 31.3%
    Labour 28.9%
    Conservative 23.7%
    Other 16.1%

    The probability of that happening in this election is much higher as long as the two largest parties continue to fight it out in the low 30’s.

    I realize it is outside the norm for those living in England to understand this, but I think those living in Scotland, Wales and Northern Ireland might understand it better because they have experienced living in a more than 2.5 party system a lot longer.

  40. @CMJ

    Ok hedging my bets somewhat but my prediction is that Labour’s election campaign will hinge on Ed Miliband’s popularity. And the Conservative’s campaign on how well the budget is received and bounce sustained.

  41. Andy Shadrack

    However, in Scotland it took a form of PR at Holyrood for even the healthily supported SNP to make an electoral breakthrough.

    Trying to get elected under FPTP is a very long business indeed – which is why it happens so rarely.

  42. @ChrisLane1945
    I play chess. I play for a club, but also online at chessworld.net

    Let me know if you want a game

  43. @LouisWVG

    Rather belatedly I have found time to look at the accuracy of your model’s projections for the March batch of Ashcroft polls (using the VI profiles you posted a couple of days ago).

    You probably won’t be surprised to learn that your rather parsimonious model doesn’t perform quite as accurately as the professional models that have everything poured into them.

    Using Ashcroft CVI figures for benchmarking, Electoral Calculus achieved a mean Euclidean Distance of 9.7 over these twelve constituencies. For Electionforecast the figure was 9.4. And for the LWVG model it turned out to be 10.8.

    I could do a paired t-test over the 12 constituencies to see whether your model was reliably less accurate. But I suspect you are not that interested in the fine details.

    Of more interest, perhaps, is the systematic bias that shows up in your projections. Over the twelve seats you reliably overstated the Tory VI (relative to CVI) and you also systematically underestimated Ukip support.

    The former bias is probably real and, interestingly, is in the opposite direction to that evident in EF and May2015. The latter may be a bit of an artefact as you provided no scores for Ukip VIs in the Scottish seats. I entered these as projections of zero, so you notched up quite a few underestimates as a result of this.

    Perhaps you would find it worthwhile to extend this kind of analysis to pick up other areas of systematic bias in your model.

  44. From the BBC

    “Comic Relief on course for £1bn

    Comic Relief is on its way to raising more than £1bn since the charity was founded 30 years ago, with more than £40 raised so far on Friday night. ”

    I suppose it’s a start but forty quid doesn’t seem a lot.

  45. While looking for something to help me understand who may vote tactically and who may not, I found this article that may help resolve many of the Scottish discussions we often have.

    http://www.britishelectionstudy.com/uncategorized/labour-need-to-tempt-not-terrify-the-voters-they-have-lost-to-the-snp/#.VQOAOtE8bIU

    It was written a month ago, perhaps Lis should have read it as it explains why “votes for the SNP help the Tories” was going to fail, as the last month has proved it has, and why tactical voting and incumbency won’t save them.

    It does however offer some suggestions to get their voters back.

  46. definite crossover….panic in labour? tory cruise to victory?

  47. AW

    congrats….this is one of the wisest things you have ever posted….

    “As I say most years, people are often too ready to assume that the budget is an opportunity for the Chancellor, when often it’s a bullet to be dodged – there are plenty of instances of budgets damaging a government and not that many of them providing a real boost. ”

    2014 budget was hyped as a “game changer” when there was barely the faintest ripple on vi….ditto last autujmn statement.

  48. Cable has said tonight it is inconceivable that there will be any arrangement between Libdems and snp.

  49. “Cable has said tonight it is inconceivable that there will be any arrangement between Libdems and snp.”

    ironic, since if there is no two party coalition which can control the house (apart from the ludicrously improbably “grand coalition”…a lib lab pact with C&S from the snp is by far the most stable 3 party coalition. on most projections such an arrangement would have a double digit majority in the commons….

    no other 3 party combination, unless you put the tories and labour together, is likely to command this.

  50. I am not sure if the following will help some people understand what happens when more than two parties are involved in a UK GE, but below please find the urls for the 1922, 1923 and 1924 UK elections:

    http://en.wikipedia.org/wiki/United_Kingdom_general_election,_1922#Results

    http://en.wikipedia.org/wiki/United_Kingdom_general_election,_1923#Results

    http://en.wikipedia.org/wiki/United_Kingdom_general_election,_1924

1 2 3 6