This week’s YouGov/Sunday Times poll is up here, with topline figures of CON 32%, LAB 33%, LDEM 7%, UKIP 15%, GRN 8%. The rest of the poll included questions on Labour and business and on intervention in Iraq.

The Conservatives are seen as having the best policies towards British business by 33% to Labour’s 19%, exactly the same split as on the economy in general. A Conservative victory at the next election is seen as being a good thing for British businesses by 44% of people, a bad thing by 23%. In contrast 43% think a Labour victory would be bad for British business, just 19% think it would be good. These questions don’t, of course, tell us whether people want a government to be good for business – when YouGov asked what the government’s attitude should be towards big business only 38% think government should primarily be supporting and helping big businesses in Britain, 49% think government should be doing more to stand up to them.

Turning to those business leaders who have criticised Ed Miliband this week, 45% of people think that the bosses of large companies should remain politically neutral, compared to 38% who think they have every right to comment on politics. There is sharp political divide on the question – Tory voters think by 59% to 31% that company bosses should intervene in politics, Labour voters think by 59% to 26% that they should keep out of politics. The idea of a CEO living in Monaco and not paying British taxes commenting on British politics goes down particularly badly, with 73% saying the intervention of Stefano Pessina is not acceptable. Nevertheless, people tend to think the criticism from business leaders is genuinely felt – 54% think business leaders are criticising Labour because they think their policies are genuinely bad for British business, 48% think they are doing do for political reasons (these includes 22% who think they are doing so for both reasons equally). 52% think that the Labour party is damaged by the comments.

YouGov also asked about intervention against Islamic State/ISIS. British air strikes against ISIS are now supported by 63% of people. YouGov asked this question very regularly last year when Britain began air strikes against ISIS, back in October 59% supported it, this is now up to 63%. 56% of people would support increasing the level of British air strikes against ISIS, but people remain opposed to putting US and British ground troops back into Iraq. 32% would support sending group troops back into Iraq, the same as when YouGov asked in October.


278 Responses to “YouGov/Sunday Times – CON 32, LAB 33, LD 7, UKIP 15, GRN 8”

1 2 3 6
  1. I expect to see the Lab vote share go up before May. probably to about 40%.

  2. 40%? Bit of wishful thinking going on there. I still expect the big two to rise though – Cons in the 33-36 range, Lab 34-38.

  3. Okay – Geek time (and apologies for the fact that this comment has been caught out by Anthony’s new thread)

    [Don’t look the other way. Admit it. Your partners and friends think you are ever so slightly weird just because you spend time on this site. So I’m talking to you! ;) ]

    Today I am going to mull over a topic that is close to the heart of psephological twitchers: namely how to identify a “crossover” : something much discussed but with confirmed sightings few and far between. It’s happened three times since the election. Swingback models assert it will happen twice again before May 7 and my own trend-based predictions posit a Green/LibDem VI crossover on March 4th followed by a Tory/Labour crossover on May 1st. UKPR posters regularly claim that one species or another of crossover is about to make an appearance.

    How can we recognise a genuine “crossover”? Spurious sighting claims may be based on one or two polls with movements well within MoE, or on graphical plots that just happen to cross over (sometimes using data from just a single polling company). But I’d like to think aloud about the criteria that should be adopted if anyone wants to make a properly substantiated claim of the crossover sighting.

    The critical distinction to make is that between plotted lines in which the crossover is consummated and those in which they merely kiss and recoil. So, for a genuine crossover there will be a situation in which the VI for Party A is statistically higher than that for Party B at Time 1, with the reverse being true at a later Time 2. It follows that a genuine incidence of crossover can only be identified in the rear-view mirror. (Even more mixed metaphors will follow below: it’s the post-prandial mead that does it..). You can’t claim a verified sighting until the initially less popular party can be demonstrated statistically to have gained the upper hand. By definition, this can only happen at some time after the event itself occurred.

    So, here’s my technical proposal for authenticating future crossover sightings. Please demur if you are minded to do so.

    (1) Use monthly batches to polls to establish the initial ascendancy of one party over another. (This can be done using simple statistical comparisons, like paired-sample t-tests.)

    (2) In cases of suspected crossover, use the same test in a later month to establish that the situation has been reversed.

    (3) Since the flip between (1) and (2) may take several months, proceed by using linear regression to fit the data for each of the parties over the changeover period.

    (4) Finally, use simultaneous equations to solve the two best fits and therefore identify the exact crossover date (admittedly with spurious precision.)

    If we followed procedures along these lines we could avoid a situation in which (more metaphors) one poster shouts “Howzat!” and others immediately take on the roles of scathing and unimpressed umpires.

    Example: Ukip/LibDem crossover in late 2012/early 2013

    Step 1: Month by month t-tests show that the LD VI was higher than that for Ukip each month up until November 2012. After that there followed a series of four months in with the VIs for the two parties did not differ statistically. Then (Step 2) in March 2013 there was a significant advantage – now in Ukip’s favour. In this case the crossover took five months to complete in a statistical sense. Applying linear regression to the Ukip and LD VIs over those 5 months, give the following equations: LD VI = 9.1 + 0.015 x [Number of days from Nov 1st] and Ukip VI = 8.28 + 0.026 x [Number of days from Nov 1st] . Solving these linear equations in Step 4 gives a Crossover date of 24th January 2013. Note that it was essential to wait until the end of March 2013 before it was possible to make a claim that crossover had occurred.

    Okay – I’ll grant you that this procedure is a bit over the top. But my point is that we need at least some kind of agreed groundrules to avoid some posters claiming crossover sighting and others bluntly denying that they were genuine. If anyone has a better idea, then I am all ears.

    PS: Over the month of January, the LibDem VI was still reliably above that for the Greens, so the crossover process has not even started in this case…

  4. @NICKP
    Blow me down, I expect the Tory vote share to go up to 40 also. What a lark if both are on 40.

  5. @Unicorn

    I assume you are refering a 95% confidence, as we are interested in both tails?

    (p<0.05 = significant)

  6. Correction

    95% confidence interval

  7. @ CMJ

    Yes – you have to use both tails, don’t you? If you put the entire rejection region in one tail you’d risk missing “kiss and recoil” encounters.

  8. That would be my view.

  9. @Rolly/NickP

    [] there is not a snowball’s chance in hell that either of the two main parties will achieve a 40% vote share in May. The Tories haven’t got close to that figure in a General Election for nigh on a quarter of a century and Labour hasn’t for 14 years. In the last two elections their combined vote share was 68% in 2005 and 66% in 2010, and it’s been heading steadily south for a generation or more.

    I suspect it will be at an all time low in a few months time and I would be astonished if either Labour or the Tories got more than 35%. My hunch is that they both may increase on the 33 v 31 current UKPR average, but not by much and we’ll see something like 34 v 32 come election time, maybe even a 33 v 33.

    If it turns out to be some toe-curling abomination of a campaign that nauseates an already disillusioned electorate, then it’s all bets off for me in terms of how poorly both Labour and the Tories will fare. That said, FPTP will mask the carnage and we’ll still see a Parliament that pretends two party politics is alive and well.

    My rider to this doleful prediction is that I still have a feeling that Miliband is the key to the whole thing. If he could somehow lift this dreary dogfight above its current wretchedness, find his voice and pitch his tent on completely different terrain, then an outright win is still there for him. He, more than other leader, has the opportunity to shift the whole terms of reference of this moribund political stalemate.

    No signs of him doing so yet, though.

  10. @Jasper

    You must have missed the polling part of your comment.

  11. My rider to this doleful prediction is that I still have a feeling that Miliband is the key to the whole thing. If he could somehow lift this dreary dogfight above its current wretchedness, find his voice and pitch his tent on completely different terrain, then an outright win is still there for him. He, more than other leader, has the opportunity to shift the whole terms of reference of this moribund political stalemate.

    Miliband simply doesn’t have the clarity, basic communication skills and eloquence to do this.

    I agree, if he had, labour could still, even with scotland, get close to a majority.

  12. @PC, aren’t Lab “close to a majority” already (if the polls are any judge)?

  13. I’d be shocked to see the Tories below 35%. They can regularly deliver VIs in that range with a conference boost or other newsworthy event, and a general election is newsworthy.

    Labour’s position is less certain because we’ve never seen them recover even temporarily from their Greenkip damage, but assuming their capacity to regain lost voters is comparable I think we can safely put them on at least 34.5%.

    @ Crossbat 11,

    If Miliband were capable of transforming the debate or his own public image, I imagine he would have done so sometime in the last two years.

    His strength, apart from holding his fractious party together, lies mainly in making the right call when he’s asked to make a sudden political judgement, eg. phone hacking, skyrocketing energy prices. He is mostly reactive. He doesn’t seem to have the ability to set the terms of debate. To be fair to him this is always quite hard to do from opposition, but he’s not doing it.

    Luckily for Labour the Government appear determined to do it for them, eg. the A&E crisis, but this is very much a Government-losing sort of election rather than an Opposition-winning sort of election and it’s hard to imagine that changing in the next few months.

  14. CB – agree to disagree.

  15. @ Unicorn,

    I take no issue with your methodology, but I’m dubious about the timescale.

    According to your method, “crossover” is only legitimate if it’s lasting, ie. if the long-term party trendlines cross and then diverge over the period of a month. It seems to me that events like conference bounces or Cameron’s original EU “veto” can produce what we should consider legitimate crossover- a real movement in the underlying voting intention, as opposed to clumped polling noise- but on a smaller timescale. Your method can’t measure these transient movements because they get folded back into the monthly trend, but I believe they are important, not least for indicating where the parties’ caps are.

    How to distinguish them numerically from clumped noise is a vexing question and not one to which I have a ready answer, but it’s one I think we at least need to consider, because by zooming out to the monthly timescale we’re sacrificing information.

  16. UNICORN

    :” But my point is that we need at least some kind of agreed groundrules to avoid some posters claiming crossover sighting and others bluntly denying that they were genuine.”

    Why do we need “groundrules” so that we all wait for two months while you tell us that crossover occured two months ago ?.

    The GE is on 7th May-so presumably from the end of this month we all just have to wait for the actual result to see if crossover has occurred.

    That’s what I’m doing anyway-but for those loyal contributors to UKPR who wish to continue calling crossover & then calling “your wrong” -I say let them carry on having fun without your “groundrules” intervening.

    Its all going to be over soon .

  17. ” A Conservative victory at the next election is seen as being a good thing for British businesses by 44% of people, a bad thing by 23%. In contrast 43% think a Labour victory would be bad for British business, just 19% think it would be good”
    ________

    Every government is good for businesses because without businesses the economy would collapse. Businesses not governments run our country.

    Onto football matters….I don’t think the protestant Muslims in Govan will be too happy with today’s result.

  18. “YouGov also asked about intervention against Islamic State/ISIS. British air strikes against ISIS are now supported by 63% of people”
    ____

    With around 900 UK nationals fighting for IS maybe we should carry out preemptive strikes a bit closer to home. It would certainly save us some money on fuel.

  19. “Its all going to be over soon .”

    A mantra for life Colin.

    I don’t find it cheers me up a lot but it does stop me ever thinking anything really matters very much.

  20. @ Speramint

    I totally agree that it would be nice to be able to detect short-term changes. But I don’t know how it could be done. One modification would be to change the batch size from monthly to (say) fortnightly or weekly. But with a smaller number of polls in each batch this could make it take longer To pick up the change. I might play around with different batch sizes

    As it happens, the Ukip/LD crossover was a rather long and drawn out process, possibly festering teo or three criss-crosses over the period in question. The SNP/Labour equivalent could have been detected much quicker.

    @ Colin – If people just want to shout at each other then they will do so as long as Anthony lets them. On the other hand, if they want to try to get to the bottoms of what is really happening then they will work together to find ways of doing this. I readily acknowledge that the detection lag makes my suggestion less than ideal. I’d be happy to step aside if someone comes up with a better process robust enough to avoid the risk of declaring twenty false crossovers for every real one. If you don’t think it is worth trying to steer away from pantomime politics then we’ll probably just have to agree to differ on this.

  21. I am not liking Allan Christie tone in his posts.

  22. “Labour’s position is less certain because we’ve never seen them recover even temporarily from their Greenkip damage, but assuming their capacity to regain lost voters is comparable I think we can safely put them on at least 34.5%.”

    But neither have we seen them facing voters with a serious choice at a general election so far.

    It seems reasonable to at least acknowledge the possibility that those “lost voters”, whilst disillusioned with Labour, were still looking for a ABT option [given that they didn’t move over directly] and that a GE may make them think:

    “Bugger: there is no option – apart from Labour – that I either prefer anyway or, more importantly, has a chance of ensuring a non Tory based government.”

    I still see both parties around 35% but with Labour having the edge.

  23. BM

    Calm down…My football mantra was reference to a joke a while back some of us engaged in. Nothing sinister.

  24. UNICORN……….Of course you don’t have to answer this question, it is purely a matter of personal interest, I am intrigued by your narrative style, could it be that you are a woman ?

  25. @ Colin

    A very nice summary, and I agree with you.

  26. ALLAN CHRISTIE……….I think BM11 was affecting an accent, if he/ she wasn’t, I’ve watched Slumdog Millionaire too often. :-)

  27. The only crossovers I can see today are the SNP/Labour (obviously in Scotland) which has been crossed over since the end of September and the other being Green/Lib/dem.

    A few months to go so still time for “crossback” or is it called swingback?

  28. Swingover?

  29. Good news for Labour today, Tony, Peter, and Alastair, have decided it’s time to come to the rescue. The Great one, fresh from his triumph as Mid-East Envoy, ( I’m thinking ISIS here ) is offering his team to Ed M. Things can only get better. :-) VI effect ?

  30. @ Unicorn and Spearmint

    Voting intentions are distictly non-continuous functions (discrete).

    This is the reason for your short term – long term dichotomy. It is a false dichotomy.. It is false because in discrete functions the transition between two points is always ex-post, it does not depend on trends as there is no way to predict the path of transition (simply because it is not observed).

    I’m quite convinced that VIs are full of trigger points – some reversible in the short term, some are not even in tong teem. This with individual variations and posing (I don’t want to see that rag (newspaper) in the house again, then buying it on the street, reading on the tablet, etc). E.g. Irreversible trigger for me any arrangement by any party with the purple. It is not measured anywhere in any kind of Bayasian model.

  31. KEN

    Och aye the noo ;-)

  32. As I browse this wonderful site, I realise that my contributions pale into insignificance when the big beast contributors, Unicorn, Laszlo, Spearmint, Statgeek, Martyn, and their ilk, post their thundering insights, H-bomb like against my pop-gun. I remain in awe, hoping to learn more, mind you, some of the geek speak confounds me. :-)

  33. Well ken thats just tony and peter fretting that they wil get the blame for a labour defeat with the consequence that the blairite project will die forever and their boy will not become leader.

    For the nats young and older -iain martin suggested on the sunday politics that in the event of lab plus snp etc > tory pluslibdem etc ..that cameron will table devo everything bar the monarchy .How do you like them onions.?

  34. ROSIEANDDAISIE…….” …….it does stop me ever thinking anything really matters very much “. Spoken like a true Nihilist. :-)

  35. @ R & D,

    But neither have we seen them facing voters with a serious choice at a general election so far.

    Nor have the Conservatives. We can’t know the true extent of ABT or ABL voting at this point- for all we know T’Other Howard’s right and the Tories will be able to scare 8% of the electorate back into their fold instead of the 2-3% we’re estimating. At this stage in the election cycle it’s an article of faith on both sides. We can’t falsify it until election day.

    What we do know is that the Tories have a proven capacity to produce blips, whereas Labour have possibly done it once, with the 2013 Conference. They’ve had fewer opportunities, but I think the absence of evidence means we have to be a bit more sceptical about their capacity to recover.

  36. As the election approaches I expect that the Tory and Labour scores should increase as people start to vote for one to stop the other (or vice versa). It becomes and Ed v Dave show.

  37. @ Ken

    Mythical beasts don’t have gender.

  38. Unicorn

    Don’t let Nessie hear you say that.

    She would be very annoyed.

  39. @ Laszlo,

    Hm…

    In the strict sense that’s obviously true because you can’t have a fractional vote, but in the broader sense I don’t think it’s quite correct. VI has discrete and continuous components. “Events, dear boy, events” are discrete but factors like the steady die-off of older, Conservative-leaning cohorts of voters are continuous (at the scale we’re looking at). And even changes triggered by events, like the collapse of the Lib Dems after the formation of the coalition, often produce what appear to be linear trends. The ex-Lib Dems didn’t all jump ship at once after the trigger, they left over time, and then there was a plateau and the losses stopped.

    I think an omniscient pollster (or possibly just Chris Huhne) could have predicted the extent of the Lib Dem losses from preexisting data. We couldn’t do it, but that’s because we didn’t know enough, not because it’s intrinsically unknowable.

  40. @ 07052015,

    For the nats young and older -iain martin suggested on the sunday politics that in the event of lab plus snp etc > tory pluslibdem etc ..that cameron will table devo everything bar the monarchy .How do you like them onions.?

    If he did it (and I doubt his backbenchers would let him) I’d almost feel sorry for Sturgeon. Back an austerity budget and a Tory Prime Minister (because of course borrowing cannot be devolved under a shared currency), or tell your supporters you turned down Devo-Max? Talk about a poisoned chalice.

  41. @ Spearmint

    I fully agree with you on the discrete and continuous components. But what are they? Do they swap places? How do we measure it? Etc.

    Thanks for the verbalisation of the issue.

    This is the reason why I abandoned my grand design of creating a non-Bayasian VI model – I bumped into the problem of “little bit pregnant”. I don’t want to bore anyone, but I think I could work out the conditionality ?s of the model, but then I met the problem of continuity and change in the same data. I was very happy because it supported my epistemological stance and despaired, because there was no way I could translate it to methodology.

    But the problem is there, and this is why Colin’s comment is so valid.

  42. Spearmint

    The monarch doesn’t set borrowing limits for Canada – or even the UK.

    Of course, taking back the property of the people that her ancestors took by force could offset the borrowing requirements by a fair amount. :-)

  43. Oldnat
    I’ve heard this said before about the monarchy. Are you talking about the Norman Conquest?

    Wasn’t George I effectively invited to become king by Parliament? And hence obviously didn’t get the crown possessions by conquest.

  44. Pete B

    George I was invited as the nearest Protestant relative of Queen Anne, not picked out of a hat!

    In any case he had expanded his Hanover domain through a series of wars.

    More importantly, a prominent No supporter, hired by the Orange Order during the referendum, claims that the Queen is directly descended from King David of Israel – and look what he did to Goliath and the Philistines!

  45. UNICON”
    @” I’d be happy to step aside”

    From what ?-have you some official position here-or are you just a contributor like the rest of us ?

    @” If you don’t think it is worth trying to steer away from pantomime politics then we’ll probably just have to agree to differ on this.”

    Unicorn , you may have read in recent times, how Health & Safety rules have not only become more pervasive, but H&S practitioners have become more intrusive. Time honoured activities providing pleasure to communities have suddenly been brought to a halt by overzealous busy bodies waving a set of H&S regulations at them despite their total irrelevance to the activities in question.

    I wonder-is the world of political polling blogs in danger of this sort of affliction? What do you think ?

  46. @Spearmint

    “What we do know is that the Tories have a proven capacity to produce blips, whereas Labour have possibly done it once, with the 2013 Conference. They’ve had fewer opportunities, but I think the absence of evidence means we have to be a bit more sceptical about their capacity to recover.”

    Depends how short-lived you like your blips, and how big too. In general terms, I don’t think you’re right and there’s a fascinating graph in the Observer today showing the polling trajectory for the five quoted parties in this Parliament. Post omn-ishambles, the Tory VI took a dip and has more or less flat-lined in the low 30’s ever since. Labour have steadily declined over the last 12 months and it’s absolutely clear that the current closeness between Labour and Tories has been due entirely to the Labour decline. I’m not sure there’s any evidence in the polls to show that the Tory VI is more elastic than Labour’s and has a greater capacity for an upsurge when stimulated. Unless, of course, I’missing something.

    I saw some comment elsewhere on here about Isobel Hardman’s guest column in today’s Observer and her reference to Crosby’s private polling and gathering Tory unity and confidence ahead of the election. This was in stark contrast to an article by Toby Helm a few pages earlier that posed the question that so agonised and traumatised poor Andrew Neil on Thursday night; “How the hell are Labour still ahead?”

    The article was a good one, nuanced and thoughtful and contained some interesting off the record quotes so beloved of political journalists. They may have all the credibility of private polls conducted by political parties, but they’re interesting nonetheless. Helm told a story of a Labour Party in some internal turmoil but developing a steely determination to win for their beleaguered and unloved leader. He likened the feeling in the party to being in the trenches, heavily sandbagged and dug in and taking in heavy artillery attack with only the occasional opportunity to return small arms fire. He quoted a Labour peer who was close to Miliband.

    When asked how it was all going, he screwed up his face. “Aaaagh,” he said with a grimace, “I think it is going to be us. It is not pretty, but I think we will do it.”

    Helm concluded:-

    “There is no point in hiding the difficulties Labour faces, but to the astonishment of many of its own troops, there remains real hope that it might still win the war.”

    So, Tory confidence growing, apparently, but some resilience and rallying around in Labour at the same time. To misquote Mark Twain, rumours of Miliband’s death much exaggerated, it would appear.

    Stop Press: – (Virgilio where art thou??)

    Reuters: President Francois Hollande’s ruling Socialists won a hotly-contested parliamentary by-election in eastern France on Sunday, narrowly beating a National Front rival after the conservative UMP candidate was eliminated in a first round.

    Hollande’s death much exaggerated too, perhaps?

    :-)

  47. @ Colin,

    I wonder-is the world of political polling blogs in danger of this sort of affliction?

    Probably not, but we do have an ongoing problem with people trying to police polling-relevant conversations they don’t like instead of just scrolling past them.

    The odds of Anthony giving Unicorn modding privileges to delete all crossover speculation that doesn’t correspond to their new definition are pretty low, so I think you can relax. Unicorn’s attempts to come up with some kind of consistent standard will in no way impede everyone else’s ability to post drivel.

  48. @Unicorn – “Over the month of January, the LibDem VI was still reliably above that for the Greens”

    That’s not how it looks here – http://en.wikipedia.org/wiki/Opinion_polling_for_the_2015_United_Kingdom_general_election#Graphical_summary – whither the discrepancy? Just too many YouGov polls in the mix there?

  49. The penalty shoot out in the African Cup is so reminiscent of Con V Lab! :-)

1 2 3 6