Sunday polls

There are three voting intention polls in the Sunday papers. The regular YouGov/Sunday Times poll has topline figures of CON 31%, LAB 38%, LDEM 9%, UKIP 13% – we’ve had a week of YouGov poll with quite low Labour leads, including a couple with leads of just one point. This seven point lead suggests they were just co-incidence and what we’re actually seeing is normal random variation around an underlying lead of 3 or 4 points (tabs are here.

Meanwhile Opinium in the Observer has topline figures of CON 29%(-3), LAB 33%(-1), LDEM 9%(+2), UKIP 20%(+2). Opinion tend to give UKIP some of their highest scores but even by those standards its a high score – the highest Opinium have shown since last summer’s 21%. We are overdue an Opinium European poll too – they said they’d be releasing one last weekend, then mid-week, but nothing yet.

Finally Survation for the Mail on Sunday have topline figures of CON 28%, LAB 34%, LDEM 10%, UKIP 20%. Survation are the other company that tends to give UKIP their highest scores but again the 20% is the highest since last summer. They also have European election voting intentions which stand at CON 21%, LAB 28%, LD 9%, UKIP 32%, GRN 5%. This seems to be becoming the broad picture of European election support (tabs here.

The rest of the YouGov and Survation polls had lots of questions about perceptions of the leaders (both found that people thought David Cameron was more “intellectually confident” than Ed Miliband. I expect it reflects general positive and negative perceptions of the two men rather than any specifics). I’m not going to drag myself over the issue with Ed Miliband’s poor ratings yet again – other than the observation that not much has changed since I last wrote about the contrast between Miliband’s ratings and Labour’s ratings.

YouGov also had some questions about rent control (people still support it) and nationalising railways (people also support it). This is one of those poll findings that are pretty consistent (I’ve seen various questions about rail nationalisation over the years and the public always seem to be supportive) yet often seem to come as a bit of a surprise to people. I suspect it’s because its one of those areas where public opinion is at odds to mainstream political debate. Outside of things like the bankers bailout nationalisation, the mainstream political debate seems to treat nationalisation as something wildly left-wing that is outside the main debate and would indicate a massive swing left, yet the public seem broadly positive towards it. Before one takes away the conclusion the public are all very left wing, one could make a similar observation about law and order, immigration or human rights policies and come to the conclusion that the public are wildly right wing – it’s more than the debate inside the party-Parliament-media sort of political universe often operates within different boundaries of what are “normal political views” from what the public actually think.

(A brief note about the differences in UKIP support between pollsters. I wrote about this last year and should probably do an updated post at some point, but the sheer quantity of ignorance and tinfoil-hattery it produces on Twitter means I should occasionally knock down some of the nonsense written. Essentially there appears to be a substantial difference between the levels of UKIP support recorded by telephone pollsters and online pollsters. Polls conducted by telephone produce lower UKIP scores, polls conducted online higher ones. We don’t know why – it could be because of an interviewer effect, or because one mode is producing more accurate samples. Differences in weighting may also be a factor – Opinium show some of the highest figures and put it down to their lack of political weighting, YouGov and Populus show lower UKIP scores than other online companies and also weight by party ID rather than past vote. What is certainly untrue is the perception that weighting by past vote automatically leads to lower UKIP scores – the method is shared by both ICM and Survation, who tend to give UKIP some of their lowest and highest scores respectively. Prompting is often cited as another cause of the gap, and it does make a difference…but Survation are the only pollster who include UKIP in their main prompt, while Opinium and ComRes online also show very high UKIP scores so this is clearly not the major factor at play. Whatever the reasons for the differences in UKIP support, they are not simple.)

146 Responses to “Sunday polls”

1 2 3
  1. Just looking to see who’d be first?

  2. Fascinating analysis Mr Wells. I do admire your capacity for neutrality.

  3. Syzygy – it was you. Woo! Yay!


    Prof C reports a new PSO referendum poll [with a non-standard question] on his blog, which apparently appears in the Sunday Mail but isn’t on the Mail/Record website. BT are claiming it’s a YouGov poll for PSO. See

    Can you shed any light on it?

  5. Barbazenzero – I couldn’t find it on their site either, but it’s generally safe to trust John Curtice.

    Progressive Polling polls are NOT YouGov polls. In the past I think YouGov have done some of their fieldwork, but that doesn’t mean they are weighted the same way as YG polls so you should compare them to the last PP poll, not to YG polls (as far as I’m aware this last one was nothing at all to do with YG and had fieldwork done elsewhere, but I couldn’t vouch for that 100%, it may be something done without my knowledge)

  6. @ AW

    There are elections in ten days (if you didn’t know!) and very possibly a high percentage will be by postal vote and very possibly a high percentage of those have already voted.

    Do any of the polling companies prompt for already voted/weight accortdingly or is there any evidence that polling accuracy improves this close to an election?

    Also it made me wonder about the BBC/Sky exit polls for the GE and how they deal with postal votes…

  7. I should add of course that only one of these polls today was relevant to elections in 10 days time…

  8. Labour & the Tories always seem to wait until the last minute to do any campaigning for the EU elections; even then, they tend to go for a general ‘support Labour’ or ‘support Conservative’ message rather than anything Europe specific.

    I still think, despite most polling evidence being to the contrary, that UKIP will do a bit worse than expected & Labour, Tory & perhaps the LDs will do a bit better than expected.

    I think this will be the outcome because of the difference which is noted between telephone & on-line polls. I do think that people who sign up for on-line polls are a little different to the rest of the voting population. I also think that Mr Nameless’s chum who switched from UKIP back to his usual Party when he voted illustrates a factor which may well have an effect on the outcome.

    I’m looking forward to the actual election result because I’m keen to know whether my assumptions are correct.

  9. Shevii – as we get closer pollsters tend to change their likelihood to vote question a bit so the 10/10 option changes from

    10/10 – Certain to vote


    10/10 – Certain to vote, or have already voted by post

    or words to that effect. Is there evidence it improves things? No, but I think it’s a reasonable thing to do.

    For the exit polls, IIRC correctly they do some backup telephone polling to get postal voters or something like that, but I’d have to check.

    Incidentally, postal votes raises an interesting legal question. It is illegal to publish figures based on people who have already voted until the polls across the whole of Europe have closed. So what about pre-election polls that include some people who have voted?

    The case was raised in 2004 when there were some regions that were all postal, so when Populus did an eve-of-poll poll, the figures for the North-West etc were *actually* exit polls, which should have been illegal. The conclusion was that’s its ok to include people who have already voted by post in polls, but not to release figures just for those who have already voted by post.


    Thanks. I thought that was the case.

    It seems odd that Prof C has included it in his poll of polls, especially as PSO don’t seem to be BPC members or publish any detailed tables.

  11. On my MOE claim in the previous comments section: it’s important to read that in context. Obviously, on a standard confidence level, Labour’s current polling AVERAGE is not within MOE of their 2010 result. The actual context in which the remark occured was Jack R s claim-

    “Don’t think Labour will be too dissatisfied with the two Westminster polls there, with one or two being a bit closer during the week.”

    – regarding two polls which had Labour on 33%. Naturally, 3a 0% is just within the MOE of a poll with a VI of 33%.

    Given that’s what I said, and given their undoubted intelligence, I struggle to understand the subject of the judicious remarks of Catmanjeff and Neil A. I suppose methodological reminders, however obvious and irrelevant, have their point.

    Roger H,

    “But even closer to their 2005 success and 60+ majority. Leads of four, five and seven points on top of a built-in electoral bias aren’t to be sniffed at.”

    (1) The 7% result was not part of that discussion, as it hadn’t been tweeted at that point.

    (2) By that logic, if Labour was on 28% and the Tories on 24%, this “wouldn’t to be sniffed at”. Yes, all of these results would deliver a Labour majority, but the idea that these results are anything other than awful for Labour in the present time astounds me. The claim that “Labour should be on 45% at this point” a few years back was silly. The claim that “Labour shouldn’t be too dissatsified with 33% at this point” doesn’t seem any better to me.

  12. Also, 2005 was a success in seats but not in terms of votes. Labour did worse than in 1979 and barely better than in 1992.

  13. Fascinating polls. There has been an air recently of ‘Labour stumbling’, but on the face of it, if taken in isolation these three polls are pretty crushing for Tories. Less than a year until the GE campaign, and they are sub 30 in two, and only just above in another.

    However, I think there are too many obvious reasons why we can’t take these polls too seriously. Just as some Lab supporters are suggesting the Euro vote might be impacting Lab VI due to a small uplift for the Greens, I also think Con’s shouldn’t get too despondent about their low scores because of UKIP. However, if they really do suffer a heavy defeat into third place on the 22nd, that could generate it’s own momentum, so there must still be concerns.

  14. Alec,

    My suspicion is that the Tories won’t improve much, if at all, before the next election, and that these polls are a good indicator of the Tories’s awful prospects in 2015. Contrary to some predictions, UKIP have benefited from the spotlight during this campaign; they will be in the spotlight moreso in the aftermath of the campaign, and then the GE will give them further attention as it does to all smaller parties.

    The Euro polls are good for Labour, albeit not as good as when it looked like they were on course for a clear win.

  15. @Anthony Wells

    The reason people support nationalisation may be because Labour voters dislike railways and Labour Governments – Attlee and Wilson in particular – were pretty fierce about cuts – Dalton’s famous phrase ‘ a mixed bag of assets ‘ comes to mind. Of course with the T&G hyolding the pursestrings Labour would hardly be anything else.

  16. Alec,

    I think the crucial difference for Labour and the Greens is that the Greens are 23 points behind Labour and UKIP are seven points ahead of the Tories.

    Voters will go back from the Greens to Labour because it won’t look like the Greens can win. Once it’s been proven that UKIP can beat the Tories in a national election, all bets may be off.

    All parties benefit from the spotlight unless it’s for extreme negative reasons (Jeremy Thorpe level allegations) so it’s no wonder both UKIP and the Greens (who have been getting a little extra airtime recently) are doing relatively well.

    I think that is the seats tally that counts; at least that is what A Level Politics students learn.

    The voting figures are affected by turn-out of voters and the impact of the third party.

    Tony Blair won three in a row on seats, I think.

  18. Sorry, UKIP are 11 points ahead of the Tories. Labour are seven points ahead of the Tories. Getting my figures mixed up.

  19. @Alec

    As I’ve said many times before, there will be all manner of tomfoolery going on in the Euro vote, less so in the local council elections I suspect, and nobody serious about these things would think that the voting behaviour we will see on May 22nd will be re-enacted, even vaguely, in the General Election in 2015. To that extent, unless you have deep concerns about who your MEP is, the election is a farce and I’d be dumbfounded if many more than 30% of the electorate bother to participate.

    However, the impact of the result on the news agenda and therefore its ability to re-set the political weather, is much more significant.

    All sorts of panic-stricken idiocy is likely to unfold, I suspect.

  20. @Anthony Wells – Rather than the populace being Left or Right, surely their views are what has always been called “populist” – it being called that because that is precisely what always tends to be popular, broadly a mix of soft economic-socialism and soft social-conservatism.

  21. I have a practical question, and I’m hoping that AW or a poster involved in the election process may know the answer please.

    When will the European election results be available?
    I know that the results will not be available at the earliest until the Europe-wide election is over, hence not until after close of polls on Sunday 25 May (as some countries are polling on Sunday I believe).

    But is it the announcement that will be delayed, or will the count also be delayed and not begin until late on Sunday night? In the latter case at least, I imagine that some results would not be available until Monday lunchtime or evening, or possibly even Tuesday.

    I should remember from 2009, but I don’t. The Electoral Commission site simply says “not earlier than 10pm on Sunday 25 May”.

  22. Alister,

    According to Wikipedia, “Most of the results of the election will be announced on Sunday 25 May, after voting has closed throughout the 28 member states of the European Union.”

    Presumably they’ll all have been counted and D’Hondt processed by then so it’s just a case of waiting to make the announcement.

  23. Bill P

    Strongly disagree, the polls this weekend are against the recent trend edging the Tories up to the mid-thirties.

    Later this year, we will see a campaign ‘6 months to save Britain’ (from Miliband) – this together with increasing economic good news and the ammunition exposing the unpleasant nature of many UKIP candidates, will turn things around.

  24. Thanks Mr Nameless.

  25. Pressman,

    You’ve (along with other press outlets) poured barrage after barrage onto UKIP and it’s hardly made a dent. In fact it may have compacted UKIP sympathisers into an ever-closer union.

    If there’s anything I’ve learned from associating with some fairly nutty social justice people on the internet, it’s that you do not win people’s support by telling them that they and the things they believe are evil. That closes off dialogue and I’m frankly shocked that nobody in your alleged organisation has seen that.

    Of course the “X Time to save the Pound” strategy was a brilliant strategy for Prime Minister Hague to focus on in 2001, and claiming Miliband to be a threat to the very existence of the nation won’t look at all hyperbolic or ridiculous.

  26. Alister –

    The legal position is that the count can take place before the close of polls elsewhere in Europe, but the results can’t be announced.

    The legal regulations explicitly set this out – they say the count can begin before or after the specified time that results can be published, and that if they take place before the returning officer has to make sure they aren’t published before the relevant time.

    I can’t remember what happens in practice either – I attended the local count at the Euros in 2009 and I think they were at some reasonable time on Sunday rather than at 10pm, but I can’t really recall it very well (was very dull)

  27. YouGov determined not to be supplanted by Populus as UKPR’s favourite pollster, I see. :p

  28. And thanks Anthony.

  29. AW – do the candidates have to wait as well?

  30. @ Bill Patrick,

    If you mean the Tories won’t recover from this weekend’s sub-30 position, that’s bonkers. This is just the annual Ukip migration. They slipped below 30% around this time last year and had bounced back to 33% within two months. (It’s Labour who should be a little uneasy about the annual Ukip migration, since last year’s migrants never migrated back and this year’s slippage leaves them only a point or two above the Tories.)

  31. Jim Jam – don’t know.

    In practice I think the local party agent at each count will be shown the local figures by the local returning officer before they are phoned into the central count for the region (or if not, will at least be able to compare the size of the piles for each party) and then be able to give that count to party headquarters, so each political party will have a fair guess of how things are going before results are actually announced.

    The votes for each counting area are also made public, but I can’t remember if that happens as it goes along or only once each region is completed.

  32. If close though agent will have to have the chance to ask for a recount and Dehont could mean close races for the final seat in some areas.
    It would be odd for the result to have to be announced before the agents could do this.
    On reflection, I reckon the agents will be given the full result but may have to sign an embargo agreement or such like.

  33. “@Anthony Wells – Rather than the populace being Left or Right, surely their views are what has always been called “populist” – it being called that because that is precisely what always tends to be popular, broadly a mix of soft economic-socialism and soft social-conservatism.”

    @ Chris Green

    I believe the party that successfully embraces both soft economic-socialism and soft social-conservatism could in the short term sweep all before them in British politics. This surely has to be the game plan for UKIP.

    However such a party would eventually need a counterpoint of soft economic and social liberalism. And that is the party I would vote for.

    Its long been my view that the more meaningful political debate in our society today is not left versus right but the divide between liberalism and all forms of state intervention.

  34. I am at a loss to predict how the seats will go in Scotland.

    On the basis of the Survation figures it would be 4 SNP and 2 Labour. But in the last few weeks it has been anything from that to 2 Lab, 2 Snp, a Tory and. UKIP!

    I had hoped for 3 SNP and with luck a Green but to be honest I think we’ll stick at two. The main reason is that the SNP’s euro campaign has been to all intents and purposes non existent.

    I am currently out a couple of mornings a week leafletting and about three evenings door knocking (partly the reason I don’t post often) but it is all for the Yes campaign.

    Not only haven’t I delivered any SNP Euro leaflets, I haven’t even seen one!


  35. I do feel that this Parliament has been a football match on a pitch with a big slope.

    Up to now, Labour has been playing down hill. They could have scored scored four possibly, but only netted two.

    Now the Conservatives are playing downhill, and have a fair wind blowing their way. Can Labour hold out to the Conservatives, including their twelfth man, the mass media?

    I really don’t know, but Ed and his team will need to be made of stern stuff.

  36. Anthony

    have you popped over to east Ukraine to do any polls?

  37. @ GRHinPorts,

    such a party would eventually need a counterpoint of soft economic and social liberalism

    Don’t we already have such a party? And they’re polling at 9%.

  38. I would have thought it would be sensible to have all elections on a Sunday or Saturday, or indeed to have the polls open on both days for a few hours, or just to have postal or electronic voting over 4 days (say). What we have is inevitably unfair to those caught out for one reason or another.

    It’s not that important anyway. Nothing much happens during the life of a parliament, whoever wins.

    Elections should be for electors, not for the entertainment of UKPR adherents.

  39. @ Spearmint

    “YouGov determined not to be supplanted by Populus as UKPR’s favourite pollster, I see. :p”

    Yes- being a gold standard only lasts as long as your last poll. Populus have some work to do on Monday’s lead if they want to retain their crown :-)

    I tend to agree with the view that polls for Labour recently should be of some concern to them. Saw the same thing during the first part of last year where the Lab drop in vote was masked by a Tory to UKIP swing with the lead unchanged. When the Tory vote swung back the lead dropped dramatically. However still a lot of reasons why the Lab vote might rise over the coming year.

  40. CatmanJeff

    That’s a good analogy, Labour have a narrow lead and already playing for the whistle hoping they can hang on. Meanwhile, the heavy artillery is coming.

  41. The interesting point in Scotland is how well the proYes parties do versus the No parties.

    I know turnout etc will make it difficult to extrapolated, but if pro-yes parties are down the we can assume that it will be a No vote as Yes is behind and the parties obv do not have a good enough organisation to get the vote out.

    And the turnout should be up if we are to assume that the referendum has energised Scottish politics – again that should be an indicator of the referendum voting.

  42. @Pressman

    Hanging on in what way? You cannot point to the run up to a completely different election, fought under an entirely different electoral system, and a turnout of no more that 35%, and extrapolate from this the likely run up to the GE.

    As for UKIP, they were 2nd in the Euros 4 years ago and then proceeded to perform poorly at the GE in 2010. There really is no reason to suspect anything other than another similar performance from UKIP in 2015.

  43. @AW,

    Survation have the official figures down as Lab 33 not 34 and, therefore, the Lab lead as 5 not 6.

  44. I seem to remember reading on a council website (Croydon perhaps) that EU Parliament election results will be announced not earlier 9.00pm on May 25th.

  45. “Whatever the reasons for the differences in UKIP support, they are not simple.)”
    Being a bit pedantic, but:
    UKIP support is what it is. What shows differences are the various methods used to attempt to measure it.

  46. *not earlier than*

  47. Billy Bob – I’ll have to check, but from memory it was going to be 9pm but then Italy moved close of polls to later moving the whole of Europe back an hour

  48. Here’s some further guidance on the count and timings:

    The count can start before 10pm, but they can’t give provisional figures to the agents at each local count until 10pm (at that point the agents will have time to digest the figures and ask for recounts before they are sent to the regional count to be added together, so even if all the counts are actually complete by 10pm (and they probably won’t be – someone will be slow!) you’re not going to get results until sometime after that)

  49. @CatmanJeff & Pressman

    And in this analogy Cameron will be playing the part of Mourinhio:

    He runs it close, but at the end he’s won b***** all and is left glowering and blaming everyone but himself :P

  50. @ Bill Patrick: “but the idea that these results are anything other than awful for Labour in the present time astounds me”

    I’m astounded that anyone can think being ahead (barring a couple of ties) in every poll continuously for over two years could be described as awful.

1 2 3