Populus’s latest poll is out and has topline figures of CON 29%(-5), LAB 40%(+1), LDEM 11%(nc), UKIP 12%(+4). Changes are from Monday and of course, while they could in theory suggest a sudden large shift from the Conservatives back to UKIP, just as likely they reflect normal random sample variation. The great benefit of high frequency polling is that we only have to wait until Populus’s next poll on Monday to find out. Full tabs are here

Meanwhile the daily YouGov poll for the Sun is much more typical of their recent polls, showing topline figures of CON 34%, LAB 40%, LDEM 10%, UKIP 11%. Full tabs here

231 Responses to “Latest YouGov and Populus polls”

1 2 3 5
  1. First

  2. second. 29%?

  3. third:
    The living standards debate and the spot checks on people from ethnic minorities at London railway stations are the big topics on BBC News.
    Very big political stories. Doreen Lawrence was on as well.

  4. (x+1)

    That’s very much not good news for DC. It’s probably just MoE, but it might knock the Tories’ confidence they’ve had over the past few weeks.

    I do have one hypothesis which I’ll have to wait for polling to bear out – The ‘nasty party’ tactics on immigration – tweeting arrest photos, the infamous vans – might be completely backfiring and, in spurring anti-immigrant rhetoric, driving some voters to the most anti-immigration of the four parties.

    We shall see.

  5. @ Ian Bailey
    ‘It seems pretty clear to me that the Labour position has softened a couple of percent towards 38’

    I think the most recent evidence suggests that there has probably been a hardening of the Labour vote back to 39/40 as Labour benefits – though to a lesser extent than the Tories – from the slippage in the UKIP vote.

  6. […]

    -It is interesting that despite the anti immigration rhetoric of Republican Congressmen most Republican voters actually are in favour of legalising illegal immigrants already in the USA.

  7. Interesting looking at the cross breaks for the Populus poll. Labour over 50% in voters 18-44. 25-34 they’re showing a 34% lead.

    It’s only in the 55+ categories the Tories lead, and they’re doing disastrously in younger demographics (17% in 25-34 – tied with the Lib Dems!)

    If I was a Conservative election strategist, this would seriously worry me. Only appealing to older voters is rubbish because in twenty years a lot of them will have died and you can’t rely on low youth turnout, especially when youth unemployment is near a million and after the anger over tuition fees.

    Unless they can remedy this, it might well be that 2015 is the last election the Tories have a chance of winning.

    Speaking of demographics, do we know why Labour do much better among women than men?


    The ‘demographic’ as people say these days, may well be a case for a less harsh approach from the Government Parties.

  9. “do we know why Labour do much better among women than men?” – Until about 15-20 years ago it used to be the other way round. It could do with the fact many are on minimum wage, and were on quite a bit less, before Labour brought it in. Labour also brought in a whole raft of maternity parent friendly policies, whilst some Conservatives even now are dead set against the minimum wage and are peeling back the maternity policies. They could also be looking at how many women MP’s they see on tv and which party they are from.

  10. RedRag – not even that far back. The data is not always consistent (MORI, for example, have sme conflicting findings – I think Roger Mortimore did a paper at EPOP which is worth looking up) but in the YouGov data that Tory female advantage only faded in early 2011.

  11. AW
    “while they could in theory suggest a sudden large shift from the Conservatives back to UKIP, just as likely they reflect normal random sample variation. The great benefit of high frequency polling is that we only have to wait until Populus’s next poll on Monday to find out. ”
    Please tell me how the next poll will tell you that.
    If it swings back, either it reflects a large shift back, or normal random variation.
    If it doesn’t swing back, it may ‘confirm’ today’s shift, or simply be another random variation in the same direction – not at all impossible.
    Given that individual shifts of up to say 3% in VI are all compatible with steady reality and random poll variation, I suggest to you that “just as likely ” is wrong, that random variation is more likely than sudden real shifts, and that more than two polls moving in the same direction are needed to reveal an underlying trend. Populus does seem to be a bit prone to ‘sudden swings’.
    That said, another 34 for the Conservatives, compared to a YG average of 31 in early July, does indicate that the Cons shares are trending slightly higher.

  12. @ Red Rag

    Also because the two main Churches started to loose (heavily) women (as they started to loose men 150 years ago).

    That’s a rather informative influence on choosing political values and with some transformation, party preference.

  13. @ Anthony Wells

    Are these female/male figures cross breaks?

  14. Dave – yes, “random variation” is indeed more likely than “sudden real shift”. I was covering my bases less I be accused of hideous Tory bias by pooh-poohing polls showing a fall in Tory support. Caught me there. Mea culpa.

    Laszlo they are, so you shouldn’t go around reading big things into individual ones!

  15. Red Rag, Anthony,

    The Con-Lab gender gap reversed somewhere before 2010 according to Peter Kellner:


    but I believe it widened in 2011 after the budget fiasco. In fact I have noticed more than once that when there are changes in VI it is women that change first.

    A very important factor is that “don’t knows” are typical twice as high for women as for men and you have to interpret the results with this in mind.

  16. The Labour vote carries on regardless, it seems set in stone, all the movement is between the Tories and UKIP, I wonder if Howell’s comments might have had an effect on Northern voters?

  17. Repost (it took a while, and it refers to today’s poll and recent polling in general).

    Has anyone noticed the Scottish cross-break? It’s a large one too (by its standards). The typical cross-break sample is 150-155, while today’s one is 213.

    Lab 41
    Con 21
    SNP 15
    Lib 14
    UKIP 6
    Green 2

    Do larger samples produce more reliable polls? SNP supporters would not agree in this case. The previous poll has a sample of 183, and the one before that a paltry 91. Add up the six most recent polls with a combined sample of 980:

    Lab 36.8
    Con 23.2
    SNP 21.7
    Lib 10.2
    UKIP 4.3
    Green 2.3

    I think that has been the source of some of the narrowing of the UK lead. Compare the MAD data for the 30 days including this most recent poll:

    Lab 40.1
    Con 21.0
    SNP 23.9
    Lib 7.6
    UKIP 3.5
    Green 1.9

    I think that Labour have been losing a little in Scotland. Where are they making up the difference though? London.

    In the first half of July, Labour dropped form low 40s to high 30s in London, but were higher elsewhere (similar, but smaller scale drop in RoS, which has a larger population). In the past two weeks, some of this has recovered, but in Scotland they have dropped from 41-42% to 37%. Something happened, that after the 22nd of July the polls drop. Something in the sampling? A politically influencing news story?

    To sum up, Labour’s lead is not solid at 39% or 40%. It is fluctuating a fair bit, but the regional changes are going unnoticed, as folk go “No change”. Let me paint you a picture of the Scottish seats:


    Lab 41
    Lib 11
    SNP 6
    Con 1

    22 polls up to 22nd of July:

    Lab 43 (+2)
    SNP 7 (+1)
    Lib 5 (-6)
    Con 4 (+3)

    Now we’ve generally seen a Lib Dem loss of six seats, and those gaining have fluctuated. However, see this:

    8 polls from 23rd July to today:

    Lab 41 (n/c)
    Lib 7 (-4)
    SNP 6 (n/c)
    Con 5 (+4)

    It’s almost as if Scotland is once again reverting to 2010 seats, with the LDs losing seats to Con. Bear in mind that the 2005 to 2010 changes were absolutely zero.

    The short, short version is, that until Labour drop below 35%, they will get about 40 seats, so while their VI has dropped, they have a good margin yet.

  18. Anthony (fpt)

    [Biased under 25 figures] should make little or no difference to the overall figures because the samples are politically weighted. If random chance gives too many Tories in the under 25s sample one day, then there will be too few somewhere else… as a whole the proportions will be the same as every other day.

    I’m not sure that is completely true. As I understand it, YouGov correct for bias based on party-id as at the last general election. But the under-25s will be less likely to have a 2010 party-id. Partly this we be an age thing, but half of them were too young to vote and presumably weren’t even part of the panel. So they are less likely to be include in an overall correction calculation.

    Along with all the other problems associated with this group (lower response, higher percentage of DKs, etc) it will mean that variation in the group will have a bigger effect than elsewhere in the sample, though ideally it should balance out over time if not on a particular day.

  19. The next Doctor Who will be announced at 7pm on Sunday


  20. Roger – nah, people who have joined the panel later still get their party ID recorded and are still weighted by it. More recent joinees are not left unweighted. We review the weighting targets each year to account for any drift in party ID amongst those who have joined more recently, and adjust the targets if necessary.

    (and higher won’t votes and won’t says means they have LESS, not MORE, of an affect on the overall result. Samples are weighted so under 25s make up 12% or so of the sample, the correct proportion of the actual population*. If a disproportionate amount of those people say they won’t vote the remainder are not weighted up to make up 12% of people with voting intentions, they just end up making up a smaller proportion of the likely voters, as they probably should)

    (*as vaguely recalled from the back of my head. Don’t take as gospel)

  21. @statgeek

    Interesting Stats, thanks.

  22. Thursday by election result at Anglesey very interesting with best ever result for PC and Conservative vote collapse…..

  23. Survation poll on East Coast Rail privatisation:


    Conservative voters want it to remain public by 48-28%, Labour voters by 63-18%, Lib Dems by 68-17%. So my question now, given that East Coast is profitable, is why are they privatising it?

  24. Mr nameless

    Your question makes no sense, how could they privatize it if it wasn’t make a profit?

  25. @mrnameless.

    ” So my question now, given that East Coast is profitable, is why are they privatising it?”

    The same question could be asked about the (part ) privatised banks.

    I think there are 2 reasons, 1) Ideology, 2) It will make the Debt figures look good if they can sell them off before 2015.

  26. It’s ironic really that the only rail line not subsidised by taxpayers is publicly owned.

  27. Can we please stop looking at crossbreaks in single polls and drawing broad demographic conclusions from them? Think of the children. Think of poor Anthony.

    Interesting poll from Populus there. I’m assuming it’s a rogue, but even accepting that it would seem to indicate their Tory baseline is a bit lower than YouGov’s despite the recent 34%. (Also, may I take this moment to commend YouGov for the clarity of their spreadsheets? Those Populus tabs are indecipherable.)

    @ Statgeek,

    Thanks for that analysis. Definitely something to keep an eye on- maybe the one Scottish Tory won’t be lonely for long!

  28. Regarding the gender gap, here’s my guess as to the reason it’s narrowing. Labour was founded as the political wing of the trade unions, an movement that was overwhelmingly male. Thus working class men would have been more directly exposed to injustices at work and become more politicised and class conscious than their wives. The transition to a largely service-based economy and the decline of trade unionism has changed this.

    Regarding the age crosstabs, I have a question for Mr Wells. I know the crosstabs are pretty small, but recent YouGov polls have consistently showed the Tories doing relatively OK among 18-24s, but the last two Populus polls have shown the Tories performing terribly among the same age group, even though they are both online polls. Are you aware of any difference in methodology between the two companies that could account for this? Could YouGov perhaps be catching more affluent, upwardly mobile young people?

  29. Colin, if Messina thinks the Tories are preferable to Labour then he’s probably a corporate, DLC-type Democrat who doesn’t share social democratic values, the kind who bequeathed Americans with NAFTA & the repeal of Glass-Steagall. He’s worked very closely over the years with Senator Max Baucus of Montana, one of the Democrats who tried very hard to undermine healthcare reform. In other words, the type Labour could do without.

    If Labour desperately want to hire some American supremo, I’d much rather they went with someone from, as Howard Dean put it, ‘The Democratic wing of the Democratic Party’ who actually shares their values


    I don’t really care what “type” he -just so long as he is successful when they hire him.

  31. @drunkenscouser,

    Yeah what Labour need is somebody like Alistair Campbell…

  32. RICH

    Yes-he shared their values I think ?

  33. After Roger Mexico’s points last week I looked into the methodology and it is clear that neither in trend nor in averages (so I don’t really understand your point about gender distribution, Anthony) in cross breaks have any meaning and the don’t knows are not comparable with other figures. Cross breaks are functions of party ID and nothing else, while DKs are unadjusted – so there is a problem because VIs are adjusted.

    The only time when cross breaks are important when YG changes allocation by party ID (at least once in every half year by the look of it, but it could be more sophisticated than this), but the effects are rather methodological (that is the adjustment is for normalising the oscillation). Basically, if radical changes happen it is normalised through adjustments in DKs and more so in cross breaks, thus smoothening the sudden change in a nice curve.

    This is not a criticism of YG, but we have to know that nothing but the headline VI are important and that adjustments take place (actually it is probably borrowed from how maths honours students are marked).

  34. Whoa, take a look at that Ynys Môn result:

    Rhun ap Iorwerth (Plaid Cymru) 12,601 / 58.2% (+16.8%)
    Tal Michael (Labour) 3,435 / 15.9% (-10.3%)
    Nathan Gill (UKIP) 3,099 / 14.3% (+14.3%)
    Neil Fairlamb (Conservative) 1,843 / 8.5% (-20.7%)
    Kathrine Jones (Socialist Labour) 348 / 1.6% (+1.6%)
    Steve Churchman (LibDems) 309 1.4% (-1.7%)
    Turnout: 42.4% (-6.3%)

    Plaid Cymru on the charge? It wouldn’t do anyone any harm for Wales to become less of a one-party state.

    And Ukip didn’t do too shabbily either. (That terrible Lib Dem result isn’t interesting; their vote has been squeezed here for a long time.)

  35. Huge swings all round then. Interesting to see Socialist Labour pop up again. Scargill’s still at it, evidently.

  36. The public sector workforce is much more female than the private sector, and attacking the public sector and its workers is an attack on some of the best jobs available to qualified, professional women – ie those most likely to vote – in many parts of the country.

    I think some sections of the commentariat need to think carefully about who and what they’re saying when they display reflexive hostility to public servants. They vote.

  37. @ Laszlo,

    The crossbreaks contain some information. We just have to understand how they’re adjusted so we’re not drawing silly conclusions from them or trying to compare apples and oranges (the DKs and the VI figures being a perfect example.)

    But I think something like Statgeek’s regional MAD figures has merit. It’s true that today’s Scotland crossbreak might be too heavily Tory, but if YouGov are consistently producing Scottish crossbreaks that are too heavily Tory they’ve screwed up. A 30-day average should mean something.

  38. Re: the YouGov poll, I think it’s quite bad for DC that almost half the country sees his party as representing only one section of society. If they make the mistake, as the Republicans have, of being seen as a party for rich old white men, they’re in danger of being boxed in.

  39. @” they’re in danger of being boxed in.”

    With a stake through their hearts ?

  40. Michael Howard’s got that covered.

  41. If the Tories continued to lag behind Labour, stuck in the early 30’s, with a bad third place in the 2014 EU elections, what are the chances of a General Election pact with UKIP?

    Personally I can’t see Nigel Farage wanting a pact, as I suspect he wants UKIP to remain a distinctly different party to the Tories. But UKIP might be prepared to do a deal with individual MP’s from all political parties who will campaign to leave the EU. I can see Cameron allowing his candidates to do this, providing they campaign on the Tory manifesto and not get involved in a UKIP campaign.

  42. @r huckle,

    What are the chances?

    Under Cameron pretty thin I would say. No love lost with ukip.

  43. Plus a formal pact with the Tories loses Ukip the North and Wales.

    Also, what’s in it for them? As I’ve said before, the Tories can’t just not put up a candidate in an election somewhere and let Ukip take the seat. Even the Lib Dems would be ashamed not to contest a GB seat in a national election in this day and age. It’s a concession of defeat- if the Tories did it they might as well just pack up their bags and go home.

    They’re going to be especially unwilling to do it in a seat that either they or Ukip might conceivably win, like Boston and Skegness. It does Ukip no good at all to have a clear run at Manchester Central.

    And Farage doesn’t believe Cameron will deliver on the referendum, because he knows perfectly well that Cameron will extract some microscopic concession from Europe, claim it as a huge victory and then campaign vigorously to stay in, at which point Ukip lose their raison d’etre and Britain is stuck in the EU for another generation.

    A pact is untenable, IMO. A Labour victory would actually be better for them, because the next Tory leader will probably be a Better Off Outer and then they’ll have the leadership a mainstream party campaigning on their side of the referendum.

  44. I think a Ukip-Boneheads pact is quite likely, though.




    “If they make the mistake, as the Republicans have, of being seen as a party for rich old white men, they’re in danger of being boxed in.

    I’m neither rich (although I do save and work so I can play) nor am I an old white man but I don’t see the Tories as being a party for the rich.

    It’s a bit like saying Labour are a party of the unemployed and those who are not willing to be transient enough to go look for better paid jobs.

  47. spearmint

    Whoa, take a look at that Ynys Môn result:

    Rhun ap Iorwerth (Plaid Cymru) 12,601 / 58.2% (+16.8%

    Do you realise just how important this win was?

    ” Despite asserting that they (Labour) never expected to win, every Welsh Government Minister joined the campaign which, if victorious, would have provided them with the overall majority at the National Assembly that has eluded them since it was set up in 1999″


  48. #stunning

  49. The question was…
    It seems to appeal to one section of society rather than to the whole country

    And 51% of the polled said …Applies most to the Conservatives

    I cannot express my own personal opinion…too partisan

1 2 3 5