ComRes have a new battleground poll out, results here, this time looking at 10 Conservative held UKIP target seats. The poll covered some seats that UKIP announced as targets last year but taking out some that no longer seem realistic, like Aylesbury, and adding some obvious ommissions like Castle Point. The overall shares of the vote with changes since 2010 were CON 39%(-7), LAB 28%(+2), UKIP 21%(+15), LDEM 5%(-10), GRN 4%(+4). UKIP have obviously advanced strongly since 2010, but remain in quite a distant third place.

It’s quite hard to know what to make of this poll. For a start, with UKIP coming from a very low base it’s quite hard to accurately predict what their best chances of a gain are – so some seats here like East Worthing and Shoreham probably aren’t in the top rank of UKIP targets. More importantly, with a poll of a group of marginal seats like this we can’t tell what the distribution of the vote is between these different seats. Most estimates are for UKIP to pick up only a couple of seats beyond their two by-election gains (which are not included in this sample), so from this poll we can’t tell whether UKIP have about 21% in all these seats (and hence wouldn’t win any of them), or are doing much better in the more promising seats like South Thanet and Thurrock and less well in the others, meaning they could win some.

I suppose all we can reasonably conclude is that UKIP don’t look like they are about to sweep the board across these seats, it’s impossible to tell from this poll whether or not they are in a stronger position in a minority of them.


484 Responses to “ComRes poll of UKIP target seats”

1 8 9 10
  1. Test

  2. @Unicorn – that’s fantastic work as usual, it’s much appreciated. I don’t dare to put forward my projections (I’m pretty sure the error is much larger than shown). I do however have a couple of questions (really more statements for comment) for you.

    1. Looking at the EF website, they have a central forecast of votes as Con 34.4%, Lab 32.6%, Lib Dems 12.5%, UKIP 10.7%, Greens 3.7%, SNP 3.6%. This is a long way away from the present polling average. Even with that, the seat projections are more or less in line with the other forecasters. Can you think of any reason for that, and how do they justify the fairly heavy “s*********”? :)

    2. I had a quick scan through the vote forecast seat-by-seat to have a look to see how far off my projections are (answer: quite a bit in some circumstances). We mostly agree on the winners in most seats – with exceptions such as Cambridge (where I don’t really take into account Huppert’s local popularity) and a few toss-up seats. My impression from the Midlands (which I still consider to a major battleground) is that the UKIP projections are a bit lower than I’d expect in Con-held seats and a bit higher in Lab-held seats. I’m also a bit surprised that the UKIP vote is so evenly distributed in the Midlands and the South – maybe I shouldn’t be!

  3. The SDLP do refer to their potential to work with SNP and Greens, but also of their traditional Labour roots. SDLP is a nationalist party but it has always considered itself a Labour party too (unlike SNP).

    You can see them as a coalition of thos who are nationalist and those who are socialist, with different people tending more to one then the other.

    In their competitionw ith Sinn Fein the idea that they are part of a PC, SNP, SDLP group is definitely quite marketable proposition.

    But I don’t think they are unhappy with the relationship they have with Labour. They have been clear they take the Labour whip but are not bound by it.

    As Labour are, in the end, the big beast of these alterantives, I don’t see SDLP explicitly rejecting their link with Labour. Moreover, that only adds fire power to those local NI Labour members – and others in the Labour party – who want Labour to stand for election in NI.

    So I think it makes sense for SDLP to talk about how they can work with Labour, SNP and PC but make no binding promises to any of these.

  4. @louiswalshvotesgreen

    I think EF have previously admitted that they don’t have much read on where Ukip are strong, because of their significant growth and lack of local data (except South Thanet and some relevant Ashcroft polls). I think they are assuming an even distribution unless there is evidence to the contrary (i.e. constituency polls).

  5. COUPER2802 & AMBER STAR

    Ladies, do please note that’s not an IFS “report” but simply another short “observation”[1] from the same David Phillips who produced the two previous observations of the 2014 GERS which he extrapolated to FFA guesstimates.

    I have no doubt that he calculated these to the very best of his considerable abilities [he got a top UCL MSc in Economics in 2009] but he’s making a procrustean bed here in going far beyond Smith and assuming that FFA or whatever the new buzzword for it is potentially achievable tomorrow whilst assuming that the total tax and spend would not change by moving functions from London to Edinburgh.

    Given Lab’s antipathy to FFA, I can’t imagine that the real negotiations could realistically get going until 2017 at the earliest, even assuming that there will be Lab minority government needing some SNP support.

    Until VE Day, it’s nothing more than base from which to argue the different LiS and SNP definitions of what “home rule” actually means in practice.

    England’s governance is clearly an issue being raised by the Cons which I would expect Lab to take very seriously if they do form the next government. That’s something which can only be sorted out via a constitutional convention, which I cannot image any of the NI, S or W MPs being against and might even prevent any of them supporting the Cons.

    [1] See http://www.ifs.org.uk/publications/7722

  6. @profhoward

    Trouble for “NI Labour” idea is that the timing for Labour making an effort in NI isn’t great, given the collapse in Labour support in the Republic. They’ve basically been supplanted by SF as the party of the left there.

  7. Little Red Rock

    “If SDLP (3), Greens (1 or 2), Plaid, (3 or 4) were enough to push Lab + LDP beyond 323”

    “That would require Labour to get about 290, which I do not expect.”

    Perhaps Labour only need 281.

    Remember that DUP (9) are open to work with Labour and they would have no problem doing so with SDLP. Jointly DUP and SDLP have much in common in their manifesto such as lower VAT four tourism (an issue relevant to Ni because of the lower vat on tourism that exists south of the irish border).

  8. Panelbase adds yet more evidence that the SNP scare tactics are truly a GAME CHANGER

    or not

    I just find it quite disappointing that serious commentators across the political spectrum expected some kind of cut through, and were already writing it up as a magnificent strategic Tory victory, dominating the air war, etc

    Of course, those of us who actually pay attention to polling saw no movement in VI in English marginals which would correlate with the rising profile of the SNP. And on the ‘issues’ polling, the obscure pontifications of a potential hung parliament are surely right down there with the EU as stuff that doesn’t actually shift votes.

    NHS, immigration, education, the economy, sure – that changes votes. Pondering obscure constitutional issues and hun pariliament dynamics is a turn off to all but the most die-hard political geeks (like me).

  9. I think Lynton Crosby (and all at Conservative Central Office) must be pinning their hopes on a swing back in the last few days/hours/as people cast their votes. Rather as must have happened in 1992

    What’s different now is the much greater number of postal votes, and I don’t think there’s as much fear about Miliband as there was about Kinnock (though this is a very recent phenomena – he’s become much more credible as the campaign has progressed)

  10. Good to see another Crosby strategy (squeeze flaky Ukipers with SNP mega scares) is another stunning success. Latest Panelbase : Tories down 2, Ukip up 1. Would you cross the road with this man?

  11. @JAMES

    You are right about the travails of Irish Labour. However NI Greens are doing well despite the similar travails of Irish Greens.

    I expect NI Labour realise that if they were to put up candidates it would be on a rather slow and small basis. Indeed their current strategy is to seek the ability only in elections involving PR (i.e. not Westminster). They seem to realise that Labour don’t want to lose the SDLP MPs by splitting a labour vote in a FPTP system.

  12. PS: Since Labour don’t stand in NI, I have noticed that NI Labour activists get up at the crack of dawn, go on the ferry, and campaign for Labour in Stranraer. So if Dumfries and Galloway has a red spot in its southwestern tip, that may be the reason.

  13. From the Newsnight blog:

    Using EF, if you take all-but-certain (95% probability or higher) victories you get a total of 488 seats. By party, this breaks down as:

    Con 228
    Lab 222
    LD 10
    SNP 25
    PC 1
    Speaker 1

    Leaving only 144 seats available. Con would have to win 98 (or Lab 104) of those seats to win a paper majority.

  14. File Panelbase under polldrums, I think.

    The “preferred coalition partners” number fits with my ongoing sense that there’s going to be a strained relationship between voters’ theoretical preferred outcome and the one they’re choosing to vote for, and that’s not going to be resolved until May 8th.

    From their 2010-now crosstabs on likely voters, pushing undecideds, the current VI split for 2010 Lib Dems is:

    Lab 29%
    Con 14%
    UKIP 22%
    Lib Dem 21%
    Green 11%
    SNP 3%

    It’s a small sample, but that direct Lib Dem – UKIP transfer of support is… interesting, and larger than the Con – UKIP transfer. I suppose you could classify it as the “something different / b*gg*r all this for a lark” vote. People are strange.

    The broader point there is that Labour has good reason to feel confident that the collapse of Lib Dem support is helping them in northern England / mid-size urban seats where they got squeezed in 2010, but that’s only going to get them so far. Likewise, the Conservatives in rural / southern England.

  15. Little Red Rock,

    I can’t see why PC, Greens would want to annoy the SNP at the outset when they are likely to be co-operating as a bloc in government. And I think Labour are serious that they won’t be offering concessions, only an ongoing process of bill-by-bill negotiation. In which case they also have nothing to gain from a specific SNP snub (as they will need reasonable relations there to get some of their bills through).

  16. Panelbase does not record much difference on a LTV test. Con gains 1% and Green losses 1%, otherwise the Lab lead would have been 4% (34 30)..

  17. ProfHoward
    “NI Labour activists get up at the crack of dawn, go on the ferry, and campaign for Labour in Stranraer”

    I wonder if Eoin Clarke is one of them? Perhaps they would be better off campaigning in a SDLP area?

  18. UNICORN
    None of the code I have seen (Firefox and IE10) does anything more than call the remote database and I have no idea how to follow it through that wormhole.

    The code itself is in one of the “include” Javascript files, but on none of the Windows browsers I have tried is it “capturable”. I would guess that’s due to a password possibly sitting there in plain sight or possibly tucked away in one of the Stylesheet files.

    ALAN

    Thanks for your offer of taking up the dropped baton!

  19. re: IFS

    Does anyone else think it a bit odd that they didn’t analyse UKIP’s proposals? I know their manifesto was “fully costed”, but it had little in the way of modeling (i.e. what would be the longer term impact of their policies, rather than the immediate cost / benefit). I know they don’t look like winning a lot of seats, but they are comfortably third in VI.

  20. @James

    I don’t know what EF do for UKIP, but one common strategy is to adjust the UKIP vote according to the voting pattern at the 2014 European Parliament elections. You may find this Wikipedia map interesting: Mapof highest polling party in each council area, 2014

  21. Holgate

    Only 15% Net 2010 LD vote gain for Lab over Cons per seat would be lower than has been typical and that Labour I expect would have been anticipating.

    Can anybody advise what the latest Churn analysis from one of our experts shows?

  22. JIM JAM

    Agreed. It’s a small sample, so high MoE, but I’d like to know more from whoever’s collating the “churnover”.

  23. April 23rd, 2015 at 3:19 pm

    tonyb

    Ashman. Best to avoid partisan posts. But here’s some facts. The economy was growing at trend, inflation was at trend. Banks blew up, starting in U.S., spreading to world. Recovery on track until 2010, then worst on record.
    ————————————————————————

    You have a valid point, I would just add one thing, up to the tax year 2009-2010 the government borrowed £160 Billion. That would have impact, on the recovery. The incoming started to cut that, therefore this would also have an impact on the recovery.

    The question is what type recovery do we want, based on borrowing or on deleveraging

    On the whole I agree on what you have stated.

  24. something that should be obvious but the media never mention it so apparently it’s not

    adding more people isn’t growth

    if you want to know what the actual growth is since 2010 you’d have to multiply the extra millions of people by some estimate of GDP per capita and then subtract that figure from the headline GDP growth figure

    the reality is all the western countries are going backwards but Britain is going backwards slightly slower than the rest because of McDoom’s tax credits (at the cost of ever increasing debt)

  25. Ashman,

    On your reply to my “conspiracy theory”.

    Of course DC would like to win and isn’t actively trying to lose. No politician ever wants to lose, but there does seem to be a lack of real hunger to get the win (contrast his energy in this campaign with John Major in 1992 for example).

    I think the reason the Eurosceptic camp have stayed quiet so far is that they do want to win so they can get the referendum which DC has promised them. They are disciplined enough to avoid opening up the internal fault lines as long as they think the party is in with a reasonable chance of forming the core of the next government. This is a fragile truce and whether it will hold to May 7th is anyone’s guess – a couple of rogue polls in a row might be enough to light the fuse.

  26. Gary O @ 16.07

    Jeremy Vine, election cowboy

    Amazingly I have spent the last 7 yrs totally oblivious to the footage of that on YouTube.

    Just watched it, amazing (ly bad)

  27. There is a new thread

  28. James

    If one side took anywhere close to that number of seats, chances are some of the seats you are considering “completely safe” would end up in play.

    They are only “safe” unconditionally (without any further information on the means and standard errors of each VI). Once you start making observations that a lot of the seats are going one way, that’s a conditional variable that probably means you have a better measurement of these means and need to reassess where the “safe line” is drawn.

    Once the exit poll is out, to an extent you have a lot more information, you have a good projection of the national VI as well as a lot of local information.

  29. @LWVG

    Even with that, the seat projections are more or less in line with the other forecasters. Can you think of any reason for that..

    For the challenger parties, their seat projections are largely decoupled from UK-wide polling figures. The Ashcroft results are dropped into the database (after modest adjustment) and then allowed to drift around with the national and regional tides. Most other models are now doing pretty much the same thing. So, whilst their regression- to- mean assumptions boost projected LD VIs and suppresses those for Ukip, it is the Ashcroft polls that have a bigger say in assigning seats to one party or another. When it comes t the Tory seat margin over Labour the seat tallies (more Tories) are in line with what your would expect from the projected VIs.

    Previous benchmarking exercises indicate that all of the models are having a tough time handling Ukip changes. This is not helped by the fact that Lord A doesn’t seem to be interested in looking all that closely at Ukip target seats. The vigour of his polling in Scotland doesn’t seem to be replicated in all other areas of the country.

  30. exileinyorks

    Ashman,

    On your reply to my “conspiracy theory”.

    Of course DC would like to win and isn’t actively trying to lose. No politician ever wants to lose, but there does seem to be a lack of real hunger to get the win (contrast his energy in this campaign with John Major in 1992 for example).

    I think the reason the Eurosceptic camp have stayed quiet so far is that they do want to win so they can get the referendum which DC has promised them. They are disciplined enough to avoid opening up the internal fault lines as long as they think the party is in with a reasonable chance of forming the core of the next government. This is a fragile truce and whether it will hold to May 7th is anyone’s guess – a couple of rogue polls in a row might be enough to light the fuse.

    ——————————————————————

    I concur DC does lack hunger, unlike Major.

    Valid point re Euro Sceptic, they are the Referendum unit, these are the people who are capable of taking on UKIP, ruthlessly, you are not even getting a squeak, out of these people in supporting their own party.
    I believe this group (Euro Sceptic) have lost the plot, it too late now. UKIP, should have been squeezed to a certain degree, it has not occurred, why keep quite, if you are the referendum party.
    The damage that will occur after the election, is far greater, for the Conservative, should the Conservative not being able to form a government.

    I care to disagree with you on this point.
    .

  31. GORDONABNERVON
    I think Lynton Crosby (and all at Conservative Central Office) must be pinning their hopes on a swing back in the last few days/hours/as people cast their votes. Rather as must have happened in 1992

    As far as I understand people now have a different understanding of what may have happened in 1992. There has been a questioning of the late swingback model. Instead it is now postulated that the polls were wrong all the way through due to systemic error.

  32. Unicorn

    I think there is a measure of allowing for MOE on the marginal polls, it’s not quite a case of overwriting existing predictions with Ashcroft polls. If there is a weird Ashcroft poll, normally it ends up somewhere between the previous prediction and the Ashcroft poll, or am I wrong?

    I noticed this with Southampton Itchen, when the original Ashcroft poll stood out like a sore thumb followed by something that looked more typical and it’s ended up somewhere between.

  33. We gat lots of talk about swingback and 1992. However one thing which is very different now is the number of postal votes cast in modern elections.

    In my ward at the last local elections almost 50% 0f the votes cast where postal votes and the vast majority of these were posted back to the polling officer with in 48 hours of landing on the doormat.

    With a higher turn out of walk in voters at a GE the proportion of votes cast as postal votes will be less than at any local election. However lets say it is 25 to30% of the total votes cast then any swing back between now and 7th May will count for nothing in a large section of the voting public.

    With this in mind and the fact that postal voting papers have been out for a few days are any of the polling organisations asking if folk have already voted. In other words doing an exit poll of postal voters. It would seem a sensible thing to do when making forecasts of final seat numbers. Or is there a law against publishing any type of exit poll even of postal votes until the ballot has closed?

  34. Are people SERIOUSLY proposing that Labour would go into coalition with the DUP? If so, it might be worth considering what happened to LibDem support when they were viewed at having betrayed their voter base by a Tory coalition. Labour might think it could never happen to them, but Scotland pretty much proves that is not the case.

1 8 9 10