The Guardian is reporting an updated version of PoliticsHome’s huge poll of marginal seats. As they did last year, PoliticsHome polled around 34,000 people in 238 marginal seats, divided into 17 groups of similar seats to make a projection of how people would vote in a general election now.

Last year it projected a Conservative majority of 146 at a time when the national polls were showing a Conservative lead of 20 points or so. This year, with the Conservative lead in the national polls down to the mid-teens, they are also doing worse in the marginals. The updated poll suggests a Conservative majority of 70. The Lib Dems are also doing significantly better. They are now projected to hold 55 seats, so down only 8.

There’s plenty of fun stuff in the full report which should be up on the PoliticsHome website tomorrow at 7am. I’ll write more then (or more probably a bit later than then. Like I get up at 7am!)


11 Responses to “PoliticsHome marginals poll”

  1. The Politics Home Map just has to be wrong. SNP winning Dundee West and Ochil & South Perthshire, I’d expect – but NE Fife? Has to be rubbish.

  2. Where’s a table on this poll? I don’t trust maps someone puts together.

  3. I reckon there might be some potentially dodgy results here (at least according to the map). Con to gain Middlesborough and Blyth Valley? (and what about Nottingham North, Warrington North and Bradford East?). At the same time Labour to hold Great Yarmouth, Chester and Wolverhampton South West?

    If I was to pick anything of big interest that might be more grounded in solid data it would be the Cumbria results as all 4 tory targets off Labour are coming up as holds (i.e. there is a consistency to the predictions that you wouldnt expect if the sample had gone funny in an individual seat). Interesting hypothesis as to why Labour may be set to do better in this region than in other marginals. After all all 4 seats (Barrow, Copeland, Workington, Carlisle) are in the Tory top 200 target list. Could it be becuase they are a mix of solid labour and solid tory support with not much swing territory? (unlike more homogenous seats in the south and midlands?) or is it because their majorities might still be slightly artificially low due to the foot and mouth issue in 2001?

    Had always thought Workington likely to stay Labour but not really any major feel for the other three.

  4. Hi JR

    You should see the tables at 10 pm Pacific Time (if I’ve worked out your time zones properly!)

  5. The Guardian report says that Hammersmith may be held by Labour but Poplar and Limehouse may fall to the Tories. Sounds a bit odd. The Tory vote is notionally about 8% higher in Hammersmith.

    Apparently Labour seats in Wolverhampton and Walsall are in danger. SNP to win only 3 seats in Scotland despite polling 30%, the Guardian reports.

  6. Andy

    UNS across the UK seems confirmed as meaningless. UNS in Scotland doesn’t seem much better.

    Within each nation, producing sets of seats with similar demographics and polling a sample within each set seems to be the only effective way forward for polling.

    Methodologies developed for a two party system with UK wide voting patterns seems quite archaic, now that the traditional fault line in UK politics has disappeared.

  7. On the 2008 map, Arundel & South Downs (on the South coast of England) appears to have been predicted as moving to the SNP with a 15% swing from Labour.

    Wrong Arundel I think…

  8. Andy – the Labour notional majority in Popular and Limehouse is 11.5%, in Hammersmith it is 13%.

  9. AW

    It seems that exactly the same seats were polled this year as last.

    So we get to know once again that the Conservatives will gain Battersea and Crawley and then see extra Conservative gains in the midlands extrapolated in unpolled constituencies.

    And South Yorkshire is again incorrectly extrapolated from West Yorkshire.

  10. Yep – absolutely the same ones, otherwise it would have been impossible to make any legitimate comparisons with last year. There’s no way some more distant prospects could have just been added onto the end of each group (or the givens knocked off).

    The only thing that could have been done was to add whole extra groups, but there really aren’t enough seats that can be considered marginals in South Yorkshire to make that worth while.

  11. First of all, congratulations on this repeat poll, Anthony. it is by far the most useful of the year. I hope it is a commercial success for you.

    In relation to South Yorkshire, it all depends what you call a “marginal”. At the levels of support when the 2008 poll was taken, \some “safe” seats were coming into the frame, but unpolled. We saw this because Cardiff South was included in the Welsh “marginals” almost to make up the numbers, but then it was discovered that it was predicted to go Tory (It isn’t this year).

    Hemsworth is one seat in South Yorkshire that whilst historically ultra-Labour has been suggested on this site as a shock result. Similarly, in West Yorkshire, Morley and Outwood where Ed Balls is standing next time. These are, however, seats where the very slight Labour recovery/Conservative sliipage recently could make all the difference. They are, I think, not regarded as marginal by PopulusHome.

    Some people, of course, will feel that if Labour are losing Cardiff South. Hemsworth and Morley the political game is decided beyond needing opinion polls.

    As Anthony says, there is a lot more to be dug out of his poll over the coming days.