Sixteen weeks to go

I’m in meetings and out and about tomorrow, so I’m doing week two’s round up tonight. The second week of 2015 and the long campaign we saw the first two phone polls of the year – Ipsos MORI’s monthly political monitor and the first weekly Ashcroft poll of the year.

YouGov/S Times (9/1/15) – CON 32%, LAB 32%, LDEM 7%, UKIP 18%, GRN 6%
Ashcroft (11/1/15) – CON 34%, LAB 28%, LDEM 8%, UKIP 16%, GRN 8%
Populus (11/1/15) – CON 32%, LAB 37%, LDEM 10%, UKIP 13%, GRN 4%
YouGov/Sun (12/1/15) – CON 32%, LAB 33%, LDEM 6%, UKIP 17%, GRN 6%
Ipsos MORI (13/1/15) – CON 33%, LAB 34%, LDEM 8%, UKIP 11%, GRN 8%
YouGov/Sun (13/1/15) – CON 32%, LAB 33%, LDEM 7%, UKIP 14%, GRN 7%
YouGov/Sun (14/1/15) – CON 32%, LAB 34%, LDEM 6%, UKIP 15%, GRN 7%
YouGov/Sun (15/1/15) – CON 32%, LAB 32%, LDEM 6%, UKIP 16%, GRN 8%

Ashcroft’s poll looks like an obvious outlier with a six point Conservative lead, most polls however clustered around a wafer thin Labour lead. The UKPR average of the latest polls now has figures of CON 33%(nc), LAB 33%(-1), LDEM 7%(-1), UKIP 15%(+1), GRN 7%(+1). Lord Ashcroft also started the year with a change to his methods, like YouGov moving to include UKIP in the main voting intention prompt.

Week 2 of the long campaign

The week started in the shadow of the Charlie Hebdo massacre, and there was some polling in the YouGov/Sunday Times poll on people’s attitudes towards free speech and religion. Around a quarter of people thought the media should not be allowed to mock or ridicule religious beliefs or figures and 18% think the media should not even be allowed to criticise or question religion. Asked about Charlie Hebdo itself 69% of people thought it was acceptable for them to publish cartoons of the Prophet Mohammed, 14% unacceptable. 63% think that other newspapers should have reprinted the cartoons, 71% that the media have an obligation to show newsworthy items even if they might upset people.

Most of the political week though has been dominated by debates over debates. David Cameron refuses to take part in debates without the Green party being invited, and the broadcasters have yet to agree to invite the Green party. Public opinion is fairly clear – two thirds of people think that Natalie Bennett should be included in the debate… but in the event that she isn’t, a majority of people think David Cameron should take part anyway.

I wrote more about the potential impact of the debates not happening yesterday, but this week the effect seems to have been an opportunity cost for the parties (time spent debating debates is time spent not talking about issues like the economy or NHS), and a spike in the Green party’s membership, which they claim has now overtaken UKIP and the Liberal Democrats.

BBC Larger and smaller parties

Following OfCom’s consultation paper last week, the BBC have released their guidance on party coverage at the election for consultation. They put UKIP in with the larger parties (Con, Lab and Lib Dem) saying “programmes must ensure that UKIP is given appropriate levels of coverage in output to which the largest parties contribute and, on some occasions, similar levels of coverage.” The Greens are under smaller parties with guidance that “programmes must ensure that the Greens are given proportionate levels of coverage in output to which the larger parties contribute, and, on occasion, similar levels of coverage, if appropriate.”


The latest forecasts from Election Forecast and May 2015 are below – both are showing a hung Parliament with Labour the largest party. Steve Fisher’s Elections Etc should be updated tomorrow.

Election Forecast – Hung Parliament, CON 277(-7), LAB 289(+8), LD 27(+1), SNP 32(-1), UKIP 3(nc)
May 2015 – Hung Parliament, CON 273(nc), LAB 280(-1), LD 24(nc), SNP 46(nc), UKIP 4(+1)

285 Responses to “Sixteen weeks to go”

1 4 5 6
  1. @AC

    Am I right that Panelbase is the most trusted pollster north of the border?

    If so (and taken alongside some other GE Scotland only polls earlier this week) I think we can safely say Labour is improving.

  2. @Unicorn

    YG only. One vested interest on my part. Can’t be arsed chasing every poll under the sun. That’s Anthony’s job. :))

  3. YG

    Con 31
    Lab 32
    LD 7
    Ukip 18
    Green 7

    YG showing the lowest Lab scores.

    Cons slipping.

  4. alec

    The thinking and conclusions seem to be stuck in a very deep groove as far as I can see.

    There is no serious basis for thinking that the Lib Dems will poll 5.5% in May or that the voters who will have deserted them since 2010, largely because they allied themselves to the Tories, will leave them, en masse in May, for the purpose of giving their votes directly to the Tories instead.

  5. RAF

    Labour could be improving but mind we have more UK polls every 3 days than we have had Scottish polls in the past 4 months yet people still can’t tell if the Tories or Labour are in the lead so like we always ask!!

    Please sir can we have some more tartan polls?

    Anyway I’m off.

    Night night

  6. @AC

    Fair point. Good night.

  7. @Allan C

    “Anyway off to bed. Taking Granny back down to the Mendips tomorrow.”

    I’ve heard of bashing the bishop but this is a euphemism I’ve never come across before.


  8. @RAF – some of us have pondered whether the falling oil price, and more specifically now the advent of job losses, might see some deflation in the mythical status of the SNP as the only defenders of the One True Scotland, but we were roundly told off on here.

    I’d really like to see further details of these polls and any ancillary questions, but I remain of the view that if a political party takes a position that is so demonstrably found to have such a weak basis in reality, then public confidence in all their other utterances is at risk.

    The counterpoint was made on here that unionist promised everything would be fine in the union, so the oil job losses will hurt them, not the SNP, but I’m less certain.

    Unionists didn’t say oil wouldn’t be volatile within the union – just that the UK would be there to help absorb the hit in the bad times.


    She lives in Frome lol

    I’m really off this time.

  10. Cross bat

    :-) :-) :-)

  11. @R&D/Alec

    I quite like old Chris Lane’s occasional eccentric contributions, but I’m not sure we should take him too seriously and, if I’ve got him right, nor do I think he would want us too either. His views seem to be totally at the mercy of the latest opinion poll and I think the Lib Dem joke is now set in stone!

  12. @Alec

    And there to leech during the good.


  13. Former Jew turned Muslim. Wow!

  14. “.. occasional eccentric contributions ………………………..

    …………..not sure we should take him too seriously”

    I don’t – but neither do I see the point in them. Are we supposed to chortle at the wit?

  15. Before we switch the lights out for the night, I thought I’d express a view on today’s clutch of polls. Opinium do tend to be severe on the Tories and kind to UKIP, but it’s still an extraordinary poll to see with less than four months to go before an election and with the starting gun more or less already fired. UKIP only 8 points behind the governing party??? YouGov and ComRes remind us, should we need to be reminded, of how poorly both the major parties are faring, with You Gov now suggesting an aggregate combined vote share of 63%. Has it ever been lower? Nothing’s impossible in politics, and a week’s a long time and all that, but it’s beyond me how some people, in the face of these persistently extraordinary polls, can still cling to pet theories forged much earlier in the Parliament. I’ve changed my mind about how I think this election will pan out at least three times in the last 18 months and I’ve done so on the basis of the polling which, if I’m honest, I’ve found astounding. It continues to astound me too.

    As for the Scottish poll, whether it’s a straw in the wind or an outlier is uncertain, but Tory strategists will be hoping and praying that the SNP aren’t faltering as we head into the home straight. Similarly, Labour will be hoping that UKIP remain strong.

    It seems to me that both the main parties, as anaemic as they appear to be, are relying on snookers and the help of a multiplicity of minor parties now.

  16. @Norbold

    I’m sure there isn’t the slightest possibility that the concealer was actually a rug rather than a prayer mat?

  17. “.. occasional eccentric contributions ………………………..
    …………..not sure we should take him too seriously”
    I don’t – but neither do I see the point in them. Are we supposed to chortle at the wit?

    Haha. Pot and kettle spring to mind. (Or is it mote & beam?)

  18. As a non purchaser of the Mail on Sunday, I’ve just dipped into their on-line version to look for the articles on Miliband mentioned by a few posters on here. I have to say I never got to read them in full, distracted as I was by a piece on why celebrities were getting disillusioned with their boob-jobs and, rather intriguingly, an article giving tips on how to detect when a woman was flirting with you. I clicked on this and it led me to some very exotic subject matter.

    I’m not easily shocked but….


  19. “Haha. Pot and kettle spring to mind. (Or is it mote & beam?)”

    I dunno Robert; I’m not the one predicting 5.5% for the Lib Dems.

  20. batty

    ” an article giving tips on how to detect when a woman was flirting with you.”

    Is it when she says things like:

    “Oi reallay loike the wy that Villa ply football owr Batty.”


  21. @RAF

    No doubt it was a prayer mat. I think it makes Ed the only Jewish Muslim who is also an atheist.

  22. Apparently there is a further YouGov poll in the Sun on Sunday!

    Lab 33 Con 31 UKIP 13 LD 7 Grn 7

  23. So two samples from the same period have Ukip on 13 and 18.

    Much variation.

  24. The MoS ‘revelations’ are courtesy of a certain Martin Winter, whose Wikipedia entry makes interesting reading:

    and as always with Doncaster you suspect there’s a lot more swept under the carpet (or possibly prayer mat). These presumably are some of the devastating shocks that Pressman promised us before Christmas. The comments on the MoS article seem to be more irate Doncastrians (of all political shades) attacking Winter than the deluge of boilerplate anti-Ed that the MoS might have been hoping for.

    So the source of the attack is pretty tainted, but anything that directs attention towards Doncaster is never going to be good for Labour.

  25. Opinium always seem to be generous to UKIP, not quite an outlier for them.

    Labour’s piddly lead still there. LD’s still unpopular, especially in Scotland. With certain parties there losing support drastically, Conservatives with solid (unreduceable?) support can hope to pick off something vulnerable. That’s if the SNP doesn’t grab them all. Fairly typical when one party’s support drops like a stone, usually when they suffer a landslide defeat there’s a bit of a feeding frenzy from everyone else.

  26. This latest YG gives Others 9%. IIRC, even if the SNP are on 50% in Scotland that is only equivalent to 4% on a UK basis. Where does the other 5% come from or have the 5 quoted VIs all been rounded up.

  27. Just looking at the Opinium tables to see what looks odd to explain their high UKIP figure vs other pollsters.

    They have to weight their retired respondents up from 302 to 506. They also had to weight up C2DE from 687 to 907.

    Just wondering what that means to margin of error. Knowing nothing about stats, perhaps someone can help explain. I often see comments like a sample size of ‘x’ will result in an MOE of 3% or 4%.

    But it would seem logical to me that if one of the weighting buckets needed to be significantly up-weighted, that would increase the margin of error?

    To take an extreme example
    – A poll of 1000 people has an MOE of 3%

    But if that poll was made up of 2 equal segments
    – segment A normally votes 70% Cons, 30% Labour
    – segment B normally votes 30% Cons, 70% Labour

    In a particular poll 1000 people were sampled
    – Segment A had a sample size of 900, result 70% Con, 30% Lab. That got downweighted to 500.
    – Segment B had a sample size of 100, result 60% Cons, 40% Lab, that gets weighted up to 500.

    But the poll is clearly wrong, as sample B was not representative as it was such a small sample.

    So the margin of error on that poll is much higher?

    Could that be part of the case with Opinium? At what point does having to do a huge amount of up weighting on a material cross break make the margin of error outside of trusted bounds?

  28. R&D
    “” an article giving tips on how to detect when a woman was flirting with you.”

    Is it when she says things like:

    “Oi reallay loike the wy that Villa ply football owr Batty.”


    I understand it is illegal to offend anyone nowadays. As a born-and-bred Brummie I must warn you that you risk being beheaded for offending me.
    P.S. In case the spooks read this thread, that was a joke!

  29. On that opinium poll, what was really odd was that they had to up-weight UKIP, from 249 to 295, so a good 20%. I virtually never see UKIP being up-weighted, so that looks really odd. And when you look at the over 55 cross break that have that as 30% Cons, 28% Lab, 27% UKIP, much higher than normal for UKIP vs other polling firms.

    That coupled with their having to materially up-weight the retired and C2DE’s (the E being pensioners) makes me suspect they have an unrepresentative sample in that retired/pensioner bucket. So they are overstating UKIP.

  30. RAF

    No Panelbase is not the most “trusted” pollster north of the border. If any one is then that position would belong to YouGov or MORI. And there hasn’t been any other specific Scottish polls since ICM in the Guardian at Christmas.

    Panelbase had the SNP only a couple of points ahead in October for Westminster when other pollsters had them surging into the 40s. Compared to that poll the SNP are up 7 and Labour down 1.

    Bottom line is that anything in the 40s is great for the SNP and guarantees them a bunch of seats. Any poll which shows a doubling of support must by definition be good news!

  31. pete

    I shall be most offended if anyone beheads me – but especially if it’s a Brummie.

  32. Alec / Chris Lane on LD losses

    Ten LD -> Lab is at the upper range of what is realistic. Check the Lab Targets on the election guide section. There are only 17 LD seats on this list, and six of those are in Scotland.

    If LDs do drop to low 20s, then that does imply that up to 20 losses to Con is conceivable. (Plus a handful to SNP)

  33. I think we have just had four polls and the Tory average is just under 31%.

    No way of knowing but I wonder if their campaigning will put as many people off as it attracts, if it is as visceral as it has been suggested that it will be.

    I’m trying to imagine what could move their VI up towards – and beyond – 2010 levels.

    Perhaps the 32% ish figure of the last year or so really is the simple reality for them.

  34. Paul H-J ,

    Labour has only a chance of 2 LD seats in Scotland , on current trends the SNP will mop up the rest bar Orkney and Shetland . Ross Skye and Lochaber is very doubtful too .

    North of Stirling , the voting system works well for the SNP

  35. @Unicorn:

    The weighting function is based on a set of zero-intercept regressions of the actual shift w.r.t the last election against the shift implied by the polls. The regression is estimated for days 365,…1 preceding the election, using data on polling and elections from 1979 onwards. We store the coefficients from the regressions, and smooth them somewhat to remove changes in the forecast induced by noise in the regression coefficients. In this sense, it’s similar to but simpler than Steve Fisher’s model, which in turn builds on this academic paper by Wlezien, Jennings, Fisher, Ford, and Pickup:

1 4 5 6