12 weeks to go

Here are this week’s polls

YouGov/S Times (6/2/15) – CON 32%, LAB 33%, LDEM 7%, UKIP 15%, GRN 8%
Opinium/Observer (6/2/15) – CON 32%, LAB 34%, LDEM 7%, UKIP 15%, GRN 8%
Ashcroft (8/2/15) – CON 34%, LAB 31%, LDEM 9%, UKIP 14%, GRN 6%
Populus (8/2/15) – CON 33%, LAB 34%, LDEM 8%, UKIP 15%, GRN 4%
YouGov/Sun (9/2/15) – CON 34%, LAB 33%, LDEM 7%, UKIP 14%, GRN 7%
YouGov/Sun (10/2/15) – CON 33%, LAB 35%, LDEM 6%, UKIP 13%, GRN 8%
Ipsos MORI/Standard (10/2/15) – CON 34%, LAB 36%, LDEM 6%, UKIP 9%, GRN 7%
YouGov/Sun (11/2/15) – CON 32%, LAB 33%, LDEM 7%, UKIP 15%, GRN 7%
YouGov/Sun (12/2/15) – CON 31%, LAB 34%, LDEM 7%, UKIP 15%, GRN 7%
Populus (12/2/15) – CON 31%, LAB 34%, LDEM 9%, UKIP 14%, GRN 6%

The big picture remains stable, with Labour and Conservatives very close. There were a couple of Conservative leads at the start of the week, but a couple of three point Labour leads at the end of the week mean the UKPR polling average continues to show a two point Labour lead – CON 32%(+1), LAB 34%(+1), LDEM 7%(-1), UKIP 14%(-1), GRN 7%(+1). Whether those two larger leads at the end of the week are anything more than normal sample noise remains to be seen. The MORI poll showing UKIP on just nine percent was their lowest since last September – I wrote on Thursday about the difficulties of actually telling what’s happening to UKIP support given changes in methodology, it looks like they might be declining slightly, but it’s hard to be sure.

Scottish and marginal polls

There were two other voting intention polls this week, a Scottish poll from TNS and a marginals poll from ComRes. The TNS poll showed a significantly smaller SNP lead than other recent Scottish polls. Survation, MORI and YouGov are all showing the SNP twenty points or more ahead; TNS produced an SNP lead of only ten points. This is presumably something to do with the methodology (TNS is a traditional face-to-face poll, rather than an online or telephone poll) as the long fieldwork for the TNS poll actually overlapped with all three of the other polls, meaning it can’t have been that opinion has changed. Meanwhile the ComRes poll showed a 4.5% swing from Con to Lab in the 40 most marginal Conservative vs Labour marginals – the equivalent of a two point lead in national polls… exactly what the national polls are currently showing!

Week 6

  • The sixth week of the long campaign started with Tristam Hunt defending comments he had made about nuns, then went through whether Labour’s women’s campaign was right to have a pink van, before settling on the rather more serious matter of tax evasion, the role of HSBC and an abortive legal threat against Ed Miliband by Lord Fink, a Tory donor with a swiss bank account. I expect we shall see some polling about tax evasion and party donors come the weekend, though possibly not about nuns.
  • UKIP launched their election campaign in the seat of Castle Point on the south coast of Essex, a potential target seat that may or may not have had UKIP’s first MP (it was never quite clear whether the defector Bob Spink formally joined UKIP or not) and where the party have formed an alliance with the Canvey Island Independents.
  • The Lib Dems launched the front page of their manifesto, showing us their main points (and their presumed demands in any coalition negotiations) will be promising to increase the personal allowance to £12,500, balance the budget fairly, invest £8bn in the NHS and provide equal mental health care, guarantee education funding and provide qualified teachers and pass “five Green Laws”.


The latest forecasts from Election Forecast, May 2015 and Elections Etc are below. The long promised seat model from the Polling Observatory team (Rob Ford, Will Jennings, Chris Wlezien and Mark Pickup) finally made its debut this week too. I don’t think it’s up on a website yet, but they put it out at the NCRM’s conference on election forecasting and tweeted it, so I’ve included it below.

Elections Etc – Hung Parliament, CON 281(-1), LAB 281(+2), LD 23(nc), SNP 41(nc), UKIP 3(nc)
Election Forecast – Hung Parliament, CON 280(-2), LAB 283(nc), LD 27(+3), SNP 37(nc), UKIP 2(nc)
May 2015 – Hung Parliament, CON 269(-1), LAB 274(+2), LD 24(-1), SNP 56(nc), UKIP 4(nc)
Polling Observatory – Hung Parliament, CON 269, LAB 293, LD 21, SNP 37, UKIP 1

209 Responses to “12 weeks to go”

1 3 4 5
  1. The last Comres poll that I can find is January 23/25th

    Changes are therefore:

    Labour +4
    Conservative +1
    UKIP -1
    LD – nc
    Green -3
    Other – nc

  2. the media thrive on drama: too close to call is not dramatic enough.

    I think that the lead up to March is critical for Cameron. If he is increasingly seen as likely to lose. then the media will get on board with that probability; that will be followed by reports of criticisms and leaks of front runners for the Tory leadership.

    And that will be followed by lower poll ratings/larger gaps. Perception is almost everything in politics. Logically it is the very reason that, at the moment, betting odds – and even “expert” opinion – have not seem that based on political realities.

    Because although it IS close, majority opinion/betting has always favoured a Tory majority of late.

  3. R&D, that won’t be the case. The cracks beneath the surface won’t emerge even if it looks likely that DC will lose; at least not until perhaps the final days by which time the mother of all political artillery will have been fired.

  4. Andy – ComRes essentially have two different series of VI polls. They do online polls for the Sunday Indy and telephone polls for the normal Indy. The two methods produce consistently different results, so we always do changes from the last ComRes poll conducted using the same method.

  5. What’s noticeable, to me, is that the drop in UKIP appears to be benefiting Labour, rather than the Conservatives. Am I reading this wrong?

  6. “Pressman

    R&D, that won’t be the case.”

    Ah… I feel my opinion is vindicated mow.

  7. meow I mean.

  8. Sorry – neow.

  9. Bugger.


    Anyway, I definitely feel vindicated but will stop trying to be precise about when that happened.

1 3 4 5