So basically polls that methodology seems to favor Labour the most are most reliable? Ashcroft normally has them roughly tied so he has a funny way of showing tory bias..

In fairness, I think Yougov are widely regarded as being the most reliable, not that I have any real knowledge of these things, I take most of it with a shovel of salt anyway.

Actually only 3 polls out of 49 in the run up to the Euro Elections had UKIP on 35% +, and 2 of those 3 were from ComRes.

In the 11 polls in the week up to the elections the UKIP average was 29.3% as against their actual result of 27.5%, an overstatement of 1.8%.

By the same token, and on the same basis, LAB was overstated by 1.6%, LD by 0.9% and Green by 0.4%. Only CON was understated - by 2.1%, enhancing the "shy Tory" theory.

Repeat that at this election from here and LAB 34 CON 34 in the polls becomes LAB 32.5 CON 36 on the day.

Similarly, if you look at the 2010 election there were 21 polls in the week up to 21 April (2 weeks before election day - the equivalent of this Thursday). They had an average CON lead of 2.4%. 4 had a LAB lead and there were two ties.

In the following week another 21 polls with an average CON lead of 4.5%. In the final week 24 polls with an average CON lead of 6.4%. Of those 8 were on polling day with an average CON lead of 7.5%. The actual CON lead was 7.1% - the lead in the polls grew by 5% in that final fortnight.

That illustrates what can happen in the next two weeks when voting with the heart is replaced by voting with the head. The question is - who will change their minds...and from which party to which. A Lib Dem revival? A minor party decline? Who benefits or suffers?

I agree I regard them highly but I find it strange to just agree to use one pollster and not take anything from other good sources too.

This is a very good analysis and why I still think the Tories will stay in power, I've changed my position though in that I think a formal coalition with the Lib Dems is more likely than a minority government (Blueisthecolour changed my mind on that).

They were also 30% in the Ashcroft poll on 9 March and 29% on 16 March (also Ashcroft). Lowest since then was 31% in a TNS poll last week.

Thank you, Radiomike, for both those posts including the two article links. It's interesting to see where the Lib Dem vote has gone (Labour, Conservative, Green, UKIP) although I don't particularly agree with Keiran Pedley's assessment of the 'shy Tories' issue because pollsters make allowances for such things after the 1992 polling debacle.

Having looked at the safer and more accurate polling for general elections since 1992, I don't buy into the late swing effect either because, even if it does exist, it's small and sporadic and it, in all probability, is not statistically significant. If the two largest parties are within a couple of hundred thousand votes of each other nationally, that other effect which occurred in 1951 and February 1974 might happen again whereby the party with the larger number of votes gets the smaller number of seats in the House of Commons. Now that is a possibility (but not necessarily a probability).

The pollsters may of learnt but still in 2010 at this point the Tories were only 3 points ahead roughly , they led on the night by 7%. Also on average in every election where polls have been used since 2010 the Tory vote share has been understated and the Lab one overstated. Do you not think that could be something to watch out for on May 7th?

Great post Mike, A+.

Is there anyway to see other constituencies data?

There's loads here: http://lordashcroftpolls.com/constituency-polls/

I'd be shocked if the Libdems hold Birmingham Yardley, which according to the polls, they shouldn't be able to.

Based on the constituency poll I'd say they have a chance.

I've actually found Ashcroft to be perfectly willing to be critical of the Tories. He's reported big Labour leads in his polls earlier this year. He's also writing a book about Cameron that DC refers to as "That Ashcroft book" and won't co-operate with.

http://static4.uk.businessinsider.co...stions%203.png

In the last Yougov poll in which Labour had a 1 point lead the people polled were asked which newspaper they read and considering Labour had a small lead it may surprise some to see the lead for the right wing newspapers

And which bar is by far the longest?

"No paper".

Sign of the times...

Wow, there is a heck of a turnaround in my constituency if you look at those poll numbers!

Seeing as we can only decide from four people (UKIP, Labour, Conservative, and Lib Dem), that would be quite an astonishing change of ways!

Although why there is even a 1% vote for green when we don't have a green person running, is just strange.

Thanks for that!

No problem , its fascinating having data like this to look through before an election.

Yeah my constituency is Crewe and Nantwich in Cheshire, and we are currently a conservative one, led by Edward Timpson, who I very rarely see around my area, nor is he easily contactable, as he is not on Twitter, nor on Facebook, so that isn't good, neither is the face that he doesn't live around here, so I won't be voting for him.

My other others are not exactly jumping out at me, nor are they plentiful in choice (three options other than him), so it looks like it is just Labour for me, as at least that candidate is on social media, and lives in the area, so is easy to find.

Mines not marginal at all, Nick boles constituency Grantham and Stamford, he'll win another landslide probably.

I'm expecting UKIP to poll a lot higher than the polls suggest. People talk of the 'shy vote' it does exist and its most likely UKIP who will benefit most from it. Too many support them but won't admit it incase they are seen as racist.

Its like the referendum, many No voters were too scared to admit it which was why the polls had the gap between Yes and No wrong until the very last moment.

As has been pointed out before, the referendum polls were almost all within the MOE.

I have seen no substantive or consistent poll swing backs for either outgoing Labour or Conservative governments. What there might very well be is post-10pm exit polls which, by their very nature, are more accurate than the pre-vote opinion polls and thus might show some different figures but that's an entirely different matter.

Whichever party ends up in second place in terms of seats in 2015 will be well placed for 2020 because there'll have been no 1983 or 1997 huge defeat.


I'd agree with that and he's certainly published individual constituency poll results that are detrimental to his own favoured party. That said, his methods might not take due account of the local popularity/incumbency factor by not mentioning the candidates' names as well as their party. I wouldn't be surprised if some Labour and Lib Dem MPs survived the predicted SNP takeover of Scotland.

Yep very good point. I think it's become more accepted to state you're going to vote UKIP but it's still not totally acceptable.

Yep, and people need to remember that these pollsters want to be as accurate as possible because their reputations are on the line. They account for many variables, because of this