Five weeks to go

A light round up this week – I’m taking it easy for the Easter weekend in advance of the long slog ahead. Here are this week’s polls (note Opinium one is a little older than some of last week’s polls – in these round ups I include polls published in the last week, even if the fieldwork is a little older.

Opinium/Observer (25/3) – CON 34%, LAB 33%, LDEM 8%, UKIP 13%, GRN 7%
YouGov/S Times (28/3) – CON 32%, LAB 36%, LDEM 8%, UKIP 13%, GRN 6%
ComRes/ITV/Mail (29/3) – CON 36%, LAB 32%, LDEM 9%, UKIP 12%, GRN 5%
Ashcroft (29/3) – CON 36%, LAB 34%, LDEM 6%, UKIP 10%, GRN 7%
Populus (29/3) – CON 34%, LAB 34%, LDEM 8%, UKIP 15%, GRN 4%
TNS BMRB (30/3) – CON 33%, LAB 32%, LDEM 8%, UKIP 16%, GRN 5%
YouGov/Sun (30/3) – CON 35%, LAB 35%, LDEM 8%, UKIP 12%, GRN 5%
YouGov/Sun (31/3) – CON 35%, LAB 36%, LDEM 7%, UKIP 12%, GRN 5%
Populus (1/4) – CON 32%, LAB 34%, LDEM 9%, UKIP 15%, GRN 5%
YouGov/Sun (1/4) – CON 36%, LAB 34%, LDEM 8%, UKIP 13%, GRN 4%
YouGov/Sun (2/4) – CON 37%, LAB 35%, LDEM 7%, UKIP 12%, GRN 5%
Panelbase (2/4) – CON 33%, LAB 33%, LDEM 7%, UKIP 17%, GRN 5%

The week started out with a bit of a false alarm as the first poll conducted after the Paxman interviews produced a four point Labour lead. It wasn’t repeated in other polling, with the rest of the week showing more typical neck-and-neck polling. The UKPR polling average is back to showing the Conservatives and Labour equal – CON 34%(+1), LAB 34%(nc), LDEM 8%(nc), UKIP 14%(nc), GRN 5%(nc).

Constituency, London, Scottish and Welsh polls

Lord Ashcroft released a new batch of constituency polls, this time revisiting some previously polled Lib Dem seats. As with his previous polling of Lib Dem seats, the key finding was variation – in some areas the Lib Dems are collapsing, in other areas they are holding up well. I wrote more about all the seats polled here, but the one that got the most attention was Sheffield Hallam, which still showed Nick Clegg’s own seat at some risk.

There were also new YouGov and ComRes polls in London, which showed 11 point and 14 point leads for Labour in the capital, and a ComRes poll in Labour held seats in Scotland, showing a swing of 19 points from Lab to SNP, the equivalent of a 16 point SNP lead in a Scottish poll. The latest YouGov poll in Wales had topline figures of CON 25%, LAB 40%, LDEM 5%, Plaid 11%, UKIP 14%, GRN 5%.

Week thirteen

The main event of the political week was the seven way leaders debate. The initial polls after the debate did not show any clear winner – different polling companies showed different politicians ahead, but broadly speaking there was a pretty even divide between Nicola Sturgeon, Ed Milband, David Cameron and Nigel Farage. It seems unlikely that a debate that produced no clear winner will have any dramatic effect upon people’s voting intentions, but we need to wait and see the evidence. While Nigel Farage didn’t win the debate, he did poll better than UKIP currently are in the polls, so it may give his party a boost. While the SNP aren’t standing across most of Britain, Nicola Sturgeon doing well could blunt the Conservative party’s warning about a Labour government reliant on SNP support.

Next week we have the close of nominations, so we’ll actually know how many candidates each party is standing. We also have the leaders debate in Scotland to come.


The latest forecasts from Election Forecast, May 2015, Elections Etc and the Guardian are below. I’ve also added the Nowcast from YouGov’s new model, which currently shows a hung Parliament like everyone else, but with more Labour seats than Conservative.

Elections Etc – Hung Parliament, CON 300(+4), LAB 258(-3), LD 20(-1), SNP 47(nc), UKIP 5(nc)
Election Forecast – Hung Parliament, CON 284(+1), LAB 276(-4), LD 26(nc), SNP 40(+2), UKIP 1(nc)
May 2015 – Hung Parliament, CON 278(+5), LAB 267(-4), LD 25(+1), SNP 54(-1), UKIP 3(-2)
Guardian – Hung Parliament, CON 276(-1), LAB 270(+1), LD 28(+3), SNP 50(-3), UKIP 4(nc)
YouGov Nowcast – Hung Parliament, CON 262, LAB 276, LD 28, SNP 56, UKIP 7

138 Responses to “Five weeks to go”

1 2 3

    You’re welcome. I’m pretty sure there’ll be a BBC NI one too, but if there is it’s not easy to find on the BBC website.

  2. I think you have all brought me here under false pretences. On the drive from Heathrow to East Preston, West Sussex, I did not see a single election sign :)

  3. @Unicorn

    have you mixed up Elections etc and Election Forecast?

  4. Smithson – Survation ENGLAND ONLY crossbreak:

    Con 34.1
    Lab 33.1
    LD 9.4
    Ukip 20.5
    Green 3

    Smithson claims that this shows a 5.2% Con-Lab swing in England.

  5. @Andy Shadrack

    That is the Tory heartland, no need for anyone to spend their precious election money on signs there ;)

  6. Where is Ashcroft’s marginal polling when we need it?

  7. @Mike49

    Elections Etc say they pick up their polls (or at least the polling average) from UKPR, but having gone for Con-Lab = 34-33 today when it’s 34 each here suggests they’re looking at the polls but not using the UKPR average

    The explanation for this discrepancy is very straightforward. Stephen Fisher routinely updates his weekly projections between 9am-10am on Friday mornings and – as we all know – Anthony posts his weekly round-up on Friday evenings. So, since up-to-date round-up figures are not available in the morning, the Electionsetc projections are always based on averages that are almost a week out of date.

  8. @Ashman

    Actually I was trying and failing to make the same point – and thought I was mocking The Mirror’s spin by using “hiding” rather than “standing” for Cameron’s physical position.

    Obviously my feeble attempt at sarcasm failed. Apologies.


    Holiday culture has changed somewhat since the 70s. Nowadays, the two English Bank Holidays wrapping the Easter Weekend are an opportunity to buy furniture at no more than 200% of its worth and to buy expensive new things for the garden to replace the ones which rotted over the winter because you had nowhere to store them.

    Electioneering proper won’t start before public offices re-open on Tuesday

    Hope you had a good flight, BTW

  10. @Spearmint

    If you were Ashcroft, would you even be in the country at this time of year?

  11. @spearmint

    Your churn report posts are on page 1 now.

    Thanks it’s an interesting and entertaining read. Recommend it to others. I had no idea Tories were benefiting that much from “Yellow Tories”.

  12. @Unicorn

    But last Friday the figures were Con 33% Lab 34%. But Elections etc have it the other way round Con 34% Lab 33%

  13. @Gary Gatter

    have you mixed up Elections etc and Election Forecast?

    Oops! I put this down to the recent proliferation of different models.

    As indicated above, my understanding is that SF uses @AW’s weighted polling averages from this very page.

  14. Richard

    Thank you for your explanation re Survation / Ukip.
    Across the board, why do you think the Ukip polling is more volatile than the other parties? It seems far more difficult to get an accurate assessment of where they are, in comparison to everyone else.

  15. @Barbazenzero

    You must introduce @Andy Shadrak to DFS.

  16. @spearmint

    re: SLD / SNP

    The Lib Dems *could* have made a breakthrough in Scotland, but they had blown it long before 2010. Their popularity in Scotland peaked in 2005, with them being the principle vehicle for anti-Iraq War sentiment, appearing to be to the left of Labour and having a popular Scottish leader. All of these factors had diminished or disappeared by 2010.

    The SNP internal coalition that you identify had started to assemble by the time of the 2007 Scottish election. With that, the Lib Dems were eclipsed.

  17. @ excileinyork

    There is that possibility I was too drunk to take in your sarcasm. :-)

  18. @unicorn

    I had a very detailed reply but web crashed and I lost the will to live. Basically, Elections etc no longer have link to this sites average on their forecasts. Maybe they no longer use the sites average or just no longer link to it as AW does not keep any historical average data here.

    Phew didn’t crash!

  19. @Spearmint

    Thanks once again for your churn report, as entertaining and enlightening as ever.

    One possible different interpretation I had was that the LD to UKIP decline seemed to mirror the LD to Con rise.

    Trying to imagine who would vote LD in 2010, then move UKIP, then move Conservative and I can only think protest voters looking for a home, so that sounds like the ABL vote up North perhaps – so possibly not going to really impact the outcome of the election. (I have been going through those Yougov nowcasts up North, and I feel for all those voters who really have no need of voting with those massive Labour majorities there – they probably feel like I feel here down South with the opposite problem)

  20. RAF
    You must introduce @Andy Shadrak to DFS.

    LOL, but what can he possibly have done to deserve that?

  21. @Spearmint

    Looking at Statgeek’s excellent regional graphs

    It looks like the Cons increase and UKIP decline is in the Midlands and Wales and up North, and they do seem to mirror each other. (more so for Midlands and Wales, up North there seems to be a Lab increase also mirroring the UKIP decline.)

  22. @Gary Gatter

    But last Friday the figures were Con 33% Lab 34%. But Elections etc have it the other way round Con 34% Lab 33%

    Hmmm..Perhaps I am out of date on this and this is one of the many changes he has made in recent weeks and months.

    This is explicitly what he did in the past, including explicit links to UKPR in the current average polls section of some of his posts:

    Date of forecast: 19 September 2014
    Days till the election: 230

    Inputted current average poll shares (from UK Polling Report)
    Con: 33%
    Lab: 35%
    LD: 8%
    Others (inc. UKIP): 24%
    – UKIP: 13%

    So, I have to admit that I no longer know exactly where his polling averages come from.

  23. @ Omnishambles,

    I had no idea Tories were benefiting that much from “Yellow Tories”.

    Nor did I, until I saw the figures. From my perspective the change is pretty shocking; LD -> Tory churn has basically been a flat line all Parliament.

    @ Richard,

    Ooh, good spot. I didn’t notice that, but you’re right, LD -> Ukip and LD -> Tory churn are almost perfectly mirrored.

    That might give us a sense of a plausible ceiling, too. LD -> Ukip churn has been bouncing around 2% since May 2013 and I’d assume most of those people are fairly committed. If the Tories are just skimming off the froth above 2% that showed up in mid-2014, they don’t have many more LD -> Ukip switchers left to absorb.

  24. @ Richard,

    Those Midlands and Wales figures are pretty grim for Labour. They need those Midlands marginals and if the Ukip vote is going over to the Tories…

  25. @Andy S

    Re signs – in this constituency they will be going up next week. Given the wind we had in the last few days, they’d be going up next week even if they’d gone up last week.

    Re Ed approval. If the Survation/Mirror story is a true reflection this is an enormous change. A month ago he was 15 points behind Cameron and there was a lot of talk about whatever happens Lab can’t win with that liability. Now his approval is only ~3 points behind Cameron so that whole story won’t wash.
    (And I have seen evidence of this on the doorstep, before the 7 way)

  26. Nichola Sturgeon

    I have started to notice that she has a unique way of speaking in which she chooses a word now and again to emphasise and she utters it in a way that is partially a laugh and partially speech. I might call it “laugh-speak”. I don’t know how else to explain it . To give it dramatic effect she always pauses before saying the word that she “laugh speaks”. Does anyone know what I mean?. It’s quite effective.

  27. Mike Smithson has put money on a spread of the Cons getting no more than a 12 seat lead (money for every Con seat fewer than that).

    Can’t say I approve of gambling – especially spread betting where losses can be huge – but MS is putting his money where his mouth is.

  28. Survation tables are now up here:

    one little thing amuses and also shows how closely matched the Parties are. Survation said The [X] Party have ruled out raising VAT if elected at the forthcoming General Election., where X was first the Conservatives and then Labour, and then asked if people believed if they would keep that promise.

    The Conservatives were disbelieved. By 51% to 49%.

    Labour were believed. By 50.2% to 49.8%.

  29. @David

    Across the board, why do you think the Ukip polling is more volatile than the other parties?

    Different methodology. AW wrote an excellent article here

    In summary, phone polling generally results in lower UKIP scores, how parties weight by party or past vote impacts (see Populus for example, they used to have the lowest UKIP scores, then they changed their party weightings in Feb and started showing UKIP scores in line with other pollsters, then UKIP dropped with other pollsters, but they continue to show high scores).

    Then you can argue prompting vs non prompting, and overeager UKIPpers more likely to respond to polls (just check comment threads online in papers to see how active they are) vs possibly shy kippers who are embarrassed to say they support them now that they are getting a bit of a racist reputation and I think its hard for the pollsters to understand what the true level of support is.

  30. Could it be that Labour’s relatively poor swing in Wales is the reason for the poor Wales+Midlands figure?

  31. Richard

    You confirm what I have been thinking.

    In the past few elections we have had plenty of evidence and talk of shy Tories. I think in this election there might be a lot of shy Ukippers but presumably it is almost impossible to calibrate.

  32. I think that Richard’s explanation of the high UKIP percentage in the Survation poll is fascinating. I wonder, if that poll was recalculated using the social class weightings used by other pollsters, what its result would show. My own feeling is that UKIP’s share is overstated by something approaching 6% and the the Conservatives should be the main beneficiary but I lack the mathematical skill to do the calculation quickly.

  33. @Spearmint

    Your entertaining and provocative churn reports are always worth waiting for. However I don’t always see what you see in your graphs.

    The Lib Dems really are breaking for the Tories. And this isn’t an artifact of a few rogue polls- it’s a consistent trend, and it’s the main source of their dramatic improvement in the past week.

    Here your summary strongly suggests that these shifting patterns go beyond being mere random fluctuations. However, if different members of the YouGov panel had been sampled, then the patterns could well have looked different. In other words – just like VIs – your X -> Y figures will have some MoE associated with them.

    The LD -> Con figure rises gently from about 3% on Jan 1 to approx 4% now. How can you be so confident that “The Lib Dems really are breaking for the Tories” rather than just showing a bit of Brownian motion?

    I also feel that sentence is a bit misleading. Your own figures show that since the last GE more 2010-LDs have ‘broken’ to Labour than to the Tories. Granted, this is not a recent shift and so is not covered by your within-2015 summary. However, if you take your full sample of 2010-LDs you will find – will you not? – that more of them have moved to Labour than to the Tories.

    Another observation I would make is that a better explanation of the Tory improvement may lie not in their success in winning over Yellow Tories, but rather in attracting back a small number of their own 2010 voters.

  34. Interesting that Milliband’s rating is rising despite the “I’ve just lost the election” headlines from the Tory press. This suggests to me that the press are not as influential as they (or politicians) think they are, or the Tories are worried about Milliband’s decreasing unpopularity (I can hardly call it rising popularity) – or both.

  35. New thread re leader ratings in Survation poll

  36. Which page / table of the Survation poll shows the (supposedly) definitive results?

    I don’t know whether to trust the 2010 (0.3 factor) weighted one, given that 2010 weightings are (in my opinion) null and void five years later, with people deserting parties for Green / UKIP / SNP (and from Lib to all).

  37. @ Barbazenzero, RAF & Guymonde

    I start work Tuesday and am just catiching up on posts, having not read anything since the debate polls came out last night (and having not slept for 48 hours until this afternoon) . So I am not sure which constitueny you are referring to. My own numbers for the first week, after averaging YouGov and Populos, and not including Survation as it is the first post debate poll:

    Conservative 34.5%
    Labour 33.3%
    UKIP 13.7%
    LD 7.7%
    National+Other 5.6%
    Green 5.2%

    I have not had a chance to review the LD weighting for 2010, but my observation from watching the debates in the Vancouver Airport was that NicK Clegg, while personally popular in the debates, LD will now face tremendous pressure in Labour and Conservative marginals.

    While I personally disagree with Farage, I think come May 7th that UkIP are primarly going to be responsible for denying the Conservatvatives the position of largest party in HoC.

    But I think one needs to wait for the second debate on April 16th, and I thought last night that the dynamics might be different with three women and two men debating.

    I noticed last night that the women were much more polite, whereas the men tried to shout ovwer each other.

    I am therefore wondering if that might change in the second debate. I am also wondering if Michelle Wood’s personal ratings had to do with her interupting the men, at one point, which Sturgeon seemed to pull off with more cache.

    Aggressive and assertive women still are looked at askance in politics, and though I disagree, I am really interested to see if there is a gender difference around the debates.

  38. Thanet

    It appears that UKIP think that they are suffering because of the turnout filter.

    They may be right. A significant proportion of UKIP voters are likely to be people who have had many opportunities to vote in previous general elections and have chosen not to.

    The pollsters appear to have decided that this makes those people less likely to vote. UKIP will say that they offer an alternative that voters have previously not had and therefore they will get votes from people who have not voted previously.

    They are both valid arguments IMO. I suppose we will find out on 8th May.

    Why are we seeing a poll almost three weeks old now?

1 2 3