Two new national polls tonight: YouGov’s daily poll in the Sun and a new entrant in the Express – Opinium Research. They have topline figures of CON 37%, LAB 30%, LDEM 16% – but more on that later. We will also have a new poll from Populus in the Times, but it is apparently going to be a marginals poll rather than their normal GB voting intention.
| 38 | 39 | 10 |
 | 8 Feb | 40 | 38 | 10 | Con +2 |
| 7 Feb | 37 | 42 | 9 | Lab +5 |
| 6 Feb | 41 | 40 | 8 | Con +1 |
| 3 Feb | 39 | 40 | 9 | Lab +1 |
| 2 Feb | 39 | 41 | 8 | Lab +2 |
| 1 Feb | 38 | 40 | 8 | Lab +2 |
| 31 Jan | 39 | 40 | 9 | Lab +1 |
| 30 Jan | 40 | 38 | 10 | Con +2 |
 | 30 Jan | 38 | 36 | 8 | Con +2 |
 | 29 Jan | 37 | 38 | 14 | Lab +1 |
| 27 Jan | 39 | 40 | 8 | Lab +1 |
| 26 Jan | 39 | 40 | 9 | Lab +1 |
| 25 Jan | 38 | 40 | 9 | Lab +2 |
| 24 Jan | 40 | 38 | 9 | Con +2 |
 | 23 Jan | 37 | 40 | 10 | Lab +3 |
 | 23 Jan | 35 | 37 | 11 | Lab +2 |
| 23 Jan | 39 | 40 | 8 | Lab +1 |
 | 23 Jan | 38 | 38 | 12 | Tie |
 | 22 Jan | 37 | 38 | 13 | Lab +1 |
 | 22 Jan | 40 | 35 | 16 | Con +5 |
| 20 Jan | 41 | 36 | 9 | Con +5 |
| 19 Jan | 38 | 38 | 11 | Tie |
| 19 Jan | 41 | 38 | 8 | Con +3 |
| 18 Jan | 40 | 39 | 7 | Con +1 |
| 17 Jan | 39 | 40 | 8 | Lab +1 |
Marginal Poll will be very interesting.
37 % for Tories is not good, but I would say 30% for Labour very bad
If the marginal pol doesn’t show at least an 8-9% Con lead, it will be most upsetting, for me anyway
Funny poll this.
37% not good for the Cons, 30% not good for Lab and 16% very bad for LD.
Who, if anyone, will be happy with this I wonder.
Three new Polls – fantastic.
However – the one already mentioned above gives the Cons 37 Lab 30 but the LD on only 16 and that leaves 17 points for ‘others’ which seems very high!!
So Labour and LD show a falling back with Cons holding up but the totals are so very interesting.
My Prediction for Yougov
Con 38
Lab 32
LD 19
These figures certaintly seem to be confirming the trend. What polling method did they use? Telephone? Face to face? Online survey? Its just that the Labour share seems to be lower than other polsters put it.
For what its worth my predictions
Marginal Poll Populus 5 % lead for Tories. Not sure what swing that woud be on 2005 , but Tories 10 seats either side of majority
You Gov 6 % for Tories
17% for others?
Well – if that’s what their results have thrown up, but surely no-one can suggest it will be anything like that high on the day.
I think adding on 3% each makes this poll a little more credible (40/33/19).
I’ve just stuck the figures into swingometer and it really illustrates how bizarre UNS looks in a Scottish context.
You’d have thought with a terrible LibDem share and a huge Others share, the SNP would be showing well. Not a bit of it. One gain. And the Libdems holding on nicely. Those Scots LibDems must have all physically upped and moved to a dozen constituencies in Scotland, leaving noone behind in the rest.
Backs up what Peter and Oldnat always tell us about the daftness of GB wide polling really.
This is stating the obvious, but this does put ‘the Others’ on 17%
Am i right in thinking they use an internet panel? Will they release their tables? Not just for this poll but for the one it is compared to?
Great news about Populus doing a marginal poll. Tories will be very keen to see another poll with a higher swing in marginal seats.
This new poll seems pretty much in line with the currant trend RE Tory lead. The others vote a bit high though.
I believe that Tories must have a lead of 7 nationally to have any chance of a majority. This of course is based on the assumption that they do better in marginals which despite what some here like to think is clearly shown in all polling on the subject so far. I think that if the lead sticks at 6 or 7 we will see a lot more marginal polling.
It’s like a convention of Mystic Megs with all these predictions
@ Neil
“I believe that Tories must have a lead of 7 nationally to have any chance of a majority”
Got to disagree there Neil. I think a solid 5% would do it, albeit a small, say 20-40 majority.
The massive support for labour in Scotland and some of the ultra safe English labour seats weigh down the percentage. A 5% national lead would equate to a residual UK lead of around 7-8%, possibly 9% sealing the Labour fate.
Only In my opinion of course.
It will be a shame if the marginals poll doesn’t have a matching national poll alongside it. Comparing one company’s marginals with other companies’ general polling would be less than adequate.
Can you clarify, by ‘they’ do you mean the new player.
So the new player has 37, 30, 16.
When can we expect the other figures?
Difficult to say whether the Opinium poll is good or bad for Labour. They report changes on last week (presumably unpublished) of Con (-2%) and Lab(+1%).
Do we know how (or whether they past-vote weight?).
Having been away for a few days, I missed the TNS/system 3 poll. From what Anthony says, it past-vote weights to actual 2005 results, so 31 for Labour was pretty good on that one – yet another piece of evidence that AR (whose low Labour score has been blamed on weighting to actual 2005 results) has some larger underlying cause for this – really only the panel left to explain it.
Also, what is the relationship between YouGov and BPIX?(I posted this on the appropriate thread, but I think everyone else has left it behind long ago).
Difficult to know without having details of their weighting. Obviously if its You Govish, its not to bad for the Tories. If its Angus Reedish, its a shocker.
Statto – YouGov do the fieldwork for BPIX (in much the same way ICM’s call centre sometimes do the fieldwork for Populus and ComRes). The question design, methodology, analysis of the data and so on are up to BPIX.
@Neil A
As you point out, Libdems in Scotland do better under FPTP than otherwise would be the case. I wonder how coy they are about ‘fair votes’ in this situation?
Why has the UKPR icon in my favourites list changed into a white g on a blue background?
@ Neil A
I wondered about the LibDems in Scotland, and apparently, the LibDems either hold the seat or are “also-rans”, unlike SNP and the Tories, who are widely spread across Scotland.
As for UK-wide polling, AR say they split the country into ~100 groups of 6-7 constituencies each, as they do to Canada, to make their polling more precise.
Personally, I’d rather see polls with 10,000 responses; if telephone and knocking-on-doors pollsters of the past managed 1,000, what’s stopping ARS/YouGov on the Internet from getting samples several times larger?
I suppose they would say “sauce for the goose is sauce for the gander”. They get a similar advantage in Cornwall, although in fairness that partly results from 20 years of Labour voters backing them to keep Tories out.
Gareth – no idea. I haven’t changed the favourite icon.
Richard Manns – two things. Firstly cost. Secondly (and perhaps more importantly) panel size. Sure you can get 10,000 people, but you can’t get that many 10,000s without the samples overlapping, especially with the less well represented groups. It’s not good to keep going back to the same people again and again too often in a short space of time (especially asking the same questions) in case it leads to a panel effect.
I suppose the answer to that AW is “quality not quantity”. Rather than 5 polls a week, how about 1 really good one every 10 days or so?!
Neil – feel free to add them together
38% for both Labour and Conservatives in the Times according to an article on their website
Ah yes, but better if the sample was taken over 2 days rather than ten. Flog your horse close to death then give it 8 days to recover….
link to that article is http://www.timesonline.co.uk/tol/comment/columnists/rachel_sylvester/article7054580.ece
I have problems with the ‘marginals’ polls, rather than the GB ones with 1000 voters. The marginals are only representative for largest party purposes when they are straight Con / Lab ones with the rest nowhere. However, the individual candidate performances over the last year or so will be presumably evident in those places, even to the diffident voter, so one would need to do a decent poll (in fact a mini election) of each to get an idea what effect the candidates are having on that particular constituency.
Neil M, Would totally agree with your posting. I do think Cons will need 7 pt lead nationally + good marginals. You will find that if the national lead increases – then they will bo better in marginals – and ofcourse the reverse.
Undecided – Flipping amazing article – loved it.
You Gov
39/34/15
How does Clegg keep his job?
LIb DEms 16-not 15
Undecided and Sue Marsh
Do you really believe they are running at 38% a piece ? Sorry, don’t believe a word of it……
Polls are all over the place – this is going to be one interesting election! I can’t wait for the day it is called.
Times article on marginals now up @ http://www.timesonline.co.uk/tol/news/politics/article7054655.ece
Populus puts the two main parties on about 38 per cent in the seats it polled. That amounts to a swing of about 6.7 per cent to the Tories from Labour since 2005 in those seats.
The poll targeted 100 Labour-held seats where the Tories came second at the last general election and which are 50 to 149 in their list of targets.
The poll excluded the 50 easiest seats for the Tories but included those the party needs to win for an overall majority.
In the key seats, Labour is still just ahead, on 38.2 per cent, down from 45.3 per cent in the 2005 election. The Tories are on 37.6 per cent, up from 31.4 per cent. This means that they should gain 97 Labour-held seats. Taking account of boundary changes, it is likely that the Conservatives would need up to 20 further seats from the Liberal Democrats and others for an overall majority.
The poll suggests that recent talk of a Tory “wobble” has affected voters’ expectations. Only 32 per cent in the marginals expect a Tory overall majority, against 43 per cent in a national poll last month. Similarly, 25 per cent now think that a Labour overall majority is most likely, against 16 per cent.