Two polls tonight – Opinium in the Observer have topline figures of CON 34%, LAB 33%, LDEM 8%, UKIP 13%, GRN 7% (tabs). All par for the course, and fieldwork was on the 24th and 25th March, so prior to the Paxman interviews on Thursday.

More intriguing is YouGov’s weekly poll for the Sunday Times. Fieldwork for this was done on Friday and Saturday, so was wholly after the Paxman interviews. Topline figures there are CON 32%, LAB 36%, LDEM 8%, UKIP 13%, GRN 6%. A four point Labour lead. I’ll write more tomorrow, but I’ll leave you with the usual caveats, it’s just one poll, it could be boost for Labour from the interviews or it could just be normal random variation, only time (and subsequent polling) will tell…


177 Responses to “Latest Opinium and YouGov polls”

1 2 3 4
  1. John TT
    Ashton-under-Lyne GE 1886. Tie. Returning Officer had casting vote. 6000 votes cast.

  2. Scottish YouGov poll in tomorrow’s SundayTtimes in Scotland rumoured to be very bad for Labour and very good indeed for SNP.

  3. @ Postage Included

    I got an interesting one from the Con candidate today, although if I didn’t know she was the Con I’d probably not have realised

    Northumbrian Scot & I were conversing about Jim Murphy’s leaflet being like this – until it dawned on me that it’s their final Constituency Report before they are no longer MPs (parliament dissolves on Monday). They’re not allowed to use their Constituency report to promote their Party.

  4. Interesting tweets

    Tom Newton Dunn [email protected] · 4 mins4 minutes ago
    Way too much of a leap to claim Lab’s 4 point lead with @YouGov tonight is down to Paxo prog, not least because this; http://sunpl.us/6015NVPL

    Toby Young [email protected] · 5 hrs5 hours ago
    Two thirds of viewers literally switched off during @Ed_Miliband’s Q&A with Paxo on Thursday night #BattleForNumber10

    I can’t read the Sun story as it is behind a paywall, but Toby seems to be saying they only changed channel when the Paxman interview started.

    I know I wanted to change the channel when that started because they started shouting at each other, same as when I accidently watch that weekly event in parliament that we are not allowed to talk about.

    But the Paxman interview was bad news for both leaders, if they watched Paxman vs Cameron, but not Paxman vs Miliband that may account for the different poll results between the instant poll (who had to watch the whole thing) and the general public, (who did not).

  5. @Spearmint

    I didn’t watch the Paxman vehicle deliberately. I like to keep my mind completely open so I can follow the party line without any niggling doubts. I’m quite sure you are wrong.

  6. @profhoward

    I don’t see the SNP position moving post-ref and the election tbh; its been odd seeing people discuss swing back when nothing of consequence has happened.

    Guess the poll has high SNP (i.e. high 40/low 50%%) bu I think its safe to say the SNP will be making considerable gains whatever.

  7. Richard. Interesting and v plausable hypothasis

  8. @two thirds of viewers switching off sounds completely ridiculous; can’t think of any programming in recent times except especially large bombs that see such a drop across programming.

    Truth is people didn’t watch Cameron in full, switched on and caught a bit then turned off for Paxman. Which only indicates the audience were more popular than the prof. interviewer!

  9. I cant spell greek (or english!)

  10. @Amber

    That doesn’t apply to the lady as she’s not an MP surely?

  11. Pete B
    Whichever candidate knocks at my door and can give your answer will win my vote. Is that an irresponsible offer?

  12. @ Postage Included

    If she’s not an MP then it definitely doesn’t apply.

  13. @ Unicorn

    You put it very eloquently my problem with those Midlands seats, and I agree.

    Thanks

  14. @Richard

    Well the broadcast media watched the whole thing and they thought EM shaded it. And EM’s margin of “victory” in the YG poll today is not consistent with people ignoring his mano a mano with Paxman as prior to the event on Thursday the voters generally thought DC better at this sort of thing.

    Again – look at the margin.

  15. @Amber
    No, this is a LD seat. As ProfH said it’s probably pandering to the anti-politics mood, but I like the idea that she’s sending a subliminal message that she’s Kipperish by going purple. Completely the wrong signal for here, in my judgement, but that’s her look-out.

  16. “Toby Young [email protected] · 5 hrs5 hours ago
    Two thirds of viewers literally switched off during @Ed_Miliband’s Q&A with Paxo on Thursday night #BattleForNumber10 ”

    Well, I switched off as soon as I saw Toby Young (though I did switch on again after I felt it was safe to assume he was gone)

  17. @JohnTT
    It’s only because of Google and Wikipedia. Though I knew there had been some very close elections I couldn’t remember the precise details. I suggest you give the doorknocker 2 minutes with his smartphone to see if he can find the answer. That’s about what it took me.

  18. According to John Plunkett in the Grauniad

    “More people were watching the non-debate at 9.45pm when Miliband started (2.8m) than were watching when Paxo pounced on Cameron (2.5m) at the top of the programme at 9pm”

    so TND & Toby Young were wrong.

  19. @RICHARD

    Re Sun switch off story.

    This story could be nonsense. In fact I read the opposite was true.

    From what I remember that was reported on the BBC website – 2.2 million were watching the Cameron part of the interview. When it was time for Miliband to be interviewed the viewing figures went up to 2.5 million.

  20. Bramley – your figures are right not mine, but same outcome.

  21. Pete B Seems as good as any other way of telling them apart!

  22. We dont how accurate this poll is – more than often than not its margin of error stuff.

    What it is an example of though is a poll that actually shapes the news agenda – its being widely reported and will be a welcome boost to the labour campaign whilst correspondingly give the tories the jitters.

    I wouldn’t be surprised if we start seeing signs of panic within the tory camp soon.

  23. Could someone (anyone!) please explain to me how these viewing figures are arrived at?

    Is it an estimate, or an accurate measure of the number of televisions tuned to that channel at the relevant time?

    If the latter, surely the number of actual viewers could be significantly higher?

    Thanks. I don’t have a television and don’t pay much attention to things like this most of the time.

  24. @ Kitsune

    Have a look at the BARB FAQs:

    http://www.barb.co.uk/resources/barb-facts/faq

  25. @Kitsune

    I believe viewing figures are still calculated by BARB http://en.m.wikipedia.org/wiki/Broadcasters%27_Audience_Research_Board

  26. By the sound of it the Sunday Times Scottish angle is not a new poll only the earlier reported BES data that 84% of the 45% those saying that they will vote SNP will certainly vote SNP. That does mean that there is up to 7% who might drift away, which is more than enough to make many of the doomed Labour seats too close to call

  27. @HEATHER PETO

    The problem with that is that a larger percentage (I don’t recall the figure, just that it was higher than that of the SNP) of Labour supporters might drift away so they could lose more than they gain. I suspect they find that somewhat less than comforting.

  28. I finally thrashed out a chart that shows the government (or opposition) lead, in way that is more sensible.

    http://www.statgeek.co.uk/polling/approval-ratings/party-lead/

    A team of clever men spent most of their Saturday evening coming up with that, messing up the spreadsheets, re-doing it all from backups and then finding the happy charts that you see before you.

    Making coloured dots is quite easy, but making them dynamic and conditional is slightly complicated, and while much head to keyboard collisions occurred, the finished product is not too bad.

    The only other issue is that now I have a slight bias towards the Conservatives, so I don’t have to re-jig my charts after May 7th. :))

  29. 07052015
    ‘You could get 2/1 Labour largest party yesterday (skybet).Thought about a treble with Rory(Masters) and Wolves for promotion to the Prem.
    Wish I had now.
    March 28th, 2015 at 10:45 pm’

    Agree, they have had a good run in the last few weeks and are starting to pick up momentum now. A very good result just before the weekend and all to play for in the last 5 weeks or so. Wolves are doing well also:-)

  30. @ Pete B

    It may be different in the UK, but in Canada election spending limits do not kick in until after the writ has been dropped.

    Up to that time individual MPs can pump out reports to their constituents to their hearts content on the taxpayers dime.

    Likewise a Party can pump out literature prior to the writ being dropped and not get caught in any spending limits either.

    Fortunately Canada Post allow their customers the ability to refuse to accept unaddressed mail, except after the writ has been dropped – but it all make great kindling for our woodstoove :)

  31. Latest yougov

    England and Wales Lab 36 -34 tory -Swing ? ,No england figure.

    Scotland Snp 45 Lab 33

    Tory lead in Midlands and Wales again.

    75 percent sure how they will vote ,20 not sure

    Lots of detail on debates and normal stuff on economy and family living standards.Some of it contradictory

    Improvements for all leaders –Miliband by 10,Clegg by 7 ,Cameron by 3.

  32. Ed Davey saying he wouldnt support coalition with tories if it included agreement on referendum.Ditto clegg and farron.Clear fault line developing .

  33. Good Morning All; cold start for Summer Time here.

    Maybe the weather has changed in the GE race. The PM seems to have started to get personal.
    Yesterday the normally fair Matthew Parris wrote about Ed Miliband havening ‘the mark of Cain’ about him; this race may get nasty.

  34. Pete B

    I must pull you up on a blatant innacuracy.

    Zaphod Beeblebrox was the Hoopiest Frood in the Galaxy.

    I can handle people misquoting Shakespeare but it shows the shocking state of our schools when they misquote Douglas Adams.

    Of course, the job of the Galactic President was not to wield power but to attract attention away from it.

    Zaphod Beeblebrox was very good at his job.

  35. Well today, March 28th in Canada, is the day we celebrate “Earth Hour”, so my spouse insisted we turn every appliance and light off we could as the object of “Earth Hour” is to see which community and province can use the least amount of energy for one hour.

    That said the cited report below, posted yesterday, underscores the need to take UKIP as a serious threat to impeding Labour in gaining the largest number of HoC seats on May 7th:

    http://www.fabians.org.uk/wp-content/uploads/2014/10/RevoltOnTheLeft-Final4.pdf

    At the end of the fourth week of March my stats averages are as follows, with the change since the beginning of March in ():

    Labour 33.6% (+.6%)
    Conservative 33.3% (-.1%)
    UKIP 13.8% (-.6%)
    Liberal Democrat 7.5%[7.7%]* (nc)
    Green Party 5.9% [5.7%]* (-.4%)
    Other (SNP/PC) 5.9% (+.5)

    *After evaluating the most recent Scottish, Welsh and London polls I have made an adjustment of .2% downwards for LD and .2% upwards for Green for the fourth week of March GB average.

    What I observe over the month is that both Labour and the nationalist parties, primarily SNP, have had a slight bounce of around .5% each, while Conservatives and Liberals have remained static and UKIP and Green have declined by around .5% each.

    There is thus no swingback to the governing parties that I can see during the month of March and in fact LD are only .2% higher than they were at the beginning of February.

    Further, the 2.6% growth in support that was seen through February for Conservative has stalled, and at the end of March the “hoped for” crossover simply has not materialized.

    An average of England-only crossbreaks for Ashcroft, Comres, Survation and Opinium in the fourth week of March reveals:

    Conservative 35.8%
    Labour 33.9%
    UKIP 14.9%
    LD 8.2%
    Green 6.4%
    Other .8%

    In 2010 the margin to Conservative in England was 5.75%, whereas at the end of March 2015 the swing back to Labour is now 4.8%. I guess that is “unclassic swingback” :)

    It is unfortunate that neither Populos nor YouGov provide England-only crossbreaks, but having determined earlier this week that the swing to Labour in London was 4% and that London represents 13.5% of the English voter turnout in 2010, I have determined that the swing to Labour outside London could be as high as 4.9%.

    Of the 81 Conservative marginals, 21 are Con-LD ones, leaving 60 potential Labour gains under a swing of 4.9% outside London and 4% in London.

    What the Fabian Society report (above), however, points out is that a bleeding of potential Labour voters to UKIP could affect the outcome in 26 Con-Labour marginals, thus reducing Labour gains from Conservative to only 34 seats at 4..9%.

    Beyond the well known example of Thurrock, there are an additional 15 seats where a swing of 2.6% or less could see these seats move to Labour, provided some of those voters do not drift from Conservative to UKIP at a rate of 3:1, projected in the Fabian report.

    Then there are an additional 10 seats where the swing needed to take the seats from Conservative ranges from 3.3% to 4.8%, where a difference of 1% or 2% in Labour support could mean taking a Conservative marginal or the Conservatives retaining it.

    Given ongoing differences of opinion between pollsters as to where UKIP values are at, and the overall tightness of the race between Labour and Conservative, the UKIP factor becomes just one more wrinkle in which the Fabian report also identifies a number of seats that Labour held in 2010 that the Conservatives could gain if more of the Labour vote slides to UKIP.

    And if this news was not bad enough, the report goes on to recognize that certain LD-Labour marginals could be impacted as well.

    So an already complex election just got a little more complex. And it is not that one should anticipate that UKIP are gong to win a whole bunch more seats, but understand that they could bleed off votes from Conservative that historically swung back and forth between Labour and Conservative.

    And in a very similar way it needs to be understood that votes that historically swung between Labour and Liberal, that used to assist Labour in a Conservative-Labour marginal, could now equally end up with the newly arriving Green candidates.

    As I have repeatedly said, this election is not a traditional two and three way election, but a four and five, and potentially six way election in Wales and Scotland.

    At the end of the third week of March the polling averages had Green at 5.1%, which I adjusted considerably upwards because I did not believe the Green value was that low. Now at the end of the fourth week the polling averages have bounced back to 5.7% or 5.9% adjusted.

    So if over the last week the nationalist and Conservative support has remained basically static, while Labour and Green support has gone up .5% and .8% respectively, and UKIP and LD support has dropped by .8% and .5%, that clearly means that there has been a flow of support from UKIP and LD to Labour and Green.

    Which means that this has been and is a multivariate election in which classic swingback will not and cannot work.

    And, if I understand it correctly, at least one of the statisticians on this list has picked up on the fact that Labour is now performing above the swingback model trajectory.

    And it will be very interesting to see if, when more national polls come out this upcoming week, they also are at variance with the GB-wide polls or in line with them.

    Finding these statistical nuances and then trying to determine if they will have any effect on the seat outcome is extremely difficult to ascertain.

    But, as I have tried to say before, an increase in the number of Green Party candidates from 335 to 554 will have an impact, as will the fact that LD might only run in 573 seats rather than 631, while UKIP, moving from 558 to 575, currently will not have any impact.

    However, the disappearance of over 500 “radical” right candidates will. Weighting to 2010, given the changed number and ideological range of candidates, has, with one exception, not been factored in by any of the pollsters, which just goes to show how difficult it is to assess the potential outcome of a multivariate election.

    Finally at the end of every week I am gong to be wondering if the movement in the polls is as result of “swingback” being slowly turned off or whether the movement between parties represents a real change.

    The fact that LD dropped .5% over the fourth week in March and Green went up .8% already has lead me to ask if that is real or is it “swingback” being turned off. So when does “swingback” start to get turned off?

  36. Sunday Herald reporting where Labour is (not) concentrating its resources in Scotland. Appears they have to prioritise because they are defending 41 seats, all of which are at some risk, and only 33 “organisers” to go round.

    The detail is not particularly surprising (i.e. Dundee West has apparently been abandoned). One interesting bit is that East Dunbartonshire, which Labour hope to gain from the Lib Dems, is having to share an organiser who is concentrating on the seat next door (Cumbernauld).

    http://www.heraldscotland.com/politics/scottish-politics/scottish-labour-accused-of-giving-up-in-key-election-marginals.121872445

  37. Zaphod’s just this guy, y’know.

  38. Appears the Sunday Times Scottish “poll” was merely them reporting on the Scottish part of the BES data. BES have written it up here:

    http://www.britishelectionstudy.com/bes-impact/the-ongoing-independence-referendum-in-scotland-implications-for-7th-may-2015/#.VRek2OFinjQ

  39. Scottish crossbreak in today’s YouGov

    SNP 45% : Lab 33% : Con 15% : LD 1% : UKIP 5% : Grn 1%

    This week’s mean of YG Scottish crossbreaks
    SNP 38% : Lab 31% : Con 17% : LD 6% : UKIP 5% : Grn 3% (Range – SNP 34-45 : Lab 27-33)

    Mean of Scottish crossbreaks in last 15 YG polls

    SNP 41% : Lab 28% : Con 18% : LD 5% : UKIP 4% : Grn 3%

    Mean of Scottish crossbreaks in last 6 Populus polls

    SNP 46% : Lab 23% : Con 15% : LD 9% : UKIP 6% : Grn 2%. (Range – SNP 45-47 : Lab 18-28)

  40. Thanks @OldNat

    So the general YouGov crossbreak movement this week looks to be an increase in variability compared to previous weeks.

  41. @ChrisLane1945

    Do you think Paris is aware of the way that St Augustine used the ‘mark of Cain’, or is he just an idiot?

  42. OLDNAT

    I am still puzzled by what is going on in the YG Scottish crossbreaks this week. I don’t really trust them.

    I am inclined to take the view that there is as yet no reliable evidence of change in Scotland.

  43. YG weekly Scottish crossbreak mean VIs Jan-March

    SNP, 40.6, 42.2, 41, 43.6, 40.6, 42.6, 42, 43, 41, 45.4, 40.8, 37.8
    Lab, 27, 26.8, 27.6, 27, 25, 27, 26.2, 24.4, 27.2, 25, 27, 30.6
    Con, 16.2, 16.3, 17.8, 19.8, 21, 16.6, 18, 16.8, 19.8, 16.4, 20, 17
    LD, 5.8, 4.3, 5.2, 3.2, 4.6, 6.2, 5.4, 4.2, 4.6, 4.4, 5.4, 5.8
    UKIP, 6.6, 5.7, 4.2, 3, 4, 3.8, 3.8, 6.6, 3, 4, 3.4, 5.4
    Grn, 4.2, 3, 3.8, 2.8, 4.2, 3, 4, 4.2, 3.8, 3.8, 2.8, 2.6

  44. Now, I don’t think you can have it both ways.
    Using crossbreaks is very dodgy, even taking averages, as we all know.

    So, if they do start to show a movement, maybe a bit SNP-> Lab swingback, we can’ then say that they are being ‘irrational’ in some way.

    If you choose to use a measure that purports to show SNP and Lab VI in Scotland, then you should stick with it, not say its behaving strangely when it shows a movement that may not be in the direction expected!!

    We need a ful scottish.
    Maybe a look at the regression analysis for the crossbreaks, modellers?

    Best
    Andyo

  45. @Kitsune

    Have a look here for how they calculate the viewing figures:

    http://www.barb.co.uk/resources/barb-facts/how-we-do-what-we-do?_s=4

  46. One thing worth noting about the YouGov Scottish sub-samples is that from Tuesday – Thursday there was a run of unusually high numbers identifying as SNP (or PC) supporters. These people were then down-weighted to the normal SNP / PC “target” score, which is based on the 2010 GE.

    The numbers identifying as SNP / PC on Friday and today are more “normal” and the SNP have gone back to more typical scores (39, 45).

  47. ANDY SHADRACK,
    hi, my understanding from previous polls suggests that the ratio of people who turn to UKIP in relation to previous conservative and Labour voters tends to be around 2-1, that is 2 previous conservative voters to every 1 Labour one. Surely this would make the marginals more secure for Labour unless I am missing something?
    The link to the Fabian Society report does suggest there could be a problem if issues are not addressed, but frankly still cannot see it and cannot see the Polling evidence for it in the Fabian Document. For example it lists Southampton Itchen as one of the places where strong UKIP showing will harm Labour, but a constituency opinion poll conducted in March 2015 shows Labour on 39%, Conservatives on 31% and UKIP on 17%. I suspect the relatively strong UKIP vote is helping Labour by taking more support from the conservatives.

  48. Andyo

    “We need a full Scottish”.

    I was hoping for one this morning. Hopefully later in the week.

  49. andyo

    “Now, I don’t think you can have it both ways.”

    “If you choose to use a measure that purports to show SNP and Lab VI in Scotland, then you should stick with it”

    I’m not sure who you were addressing that comment to.

    Initially, I chose to average the YG Scottish crossbreaks because YG changed their methodology in Jan to weight their Scottish crossbreaks (at least partially) to reflect Scottish, as opposed to just a component of GB opinion.

    I have stuck to it, in order to create a time sequence that would indicate variations in YG’s sampling that might also be replicated in “Full Scottish” polls, and

    Having set up the spreadsheet, I’ll continue with it through to the election. Any conclusions that people may draw from the data is their concern!

  50. James

    I noted the weighting factor for SNP/PC point in the middle of the week.

    However, when I checked a sample from a longer time period, there was no obvious correlation between that and higher/lower SNP VI.

1 2 3 4