Politicshome, which officially launches today, has up as Breaking news that the lastest poll (presumably YouGov for the Evening Standard) has Boris Johnson leading Ken Livingstone by 13 points. I’ll post properly once I’ve seen the poll.

UPDATE: The new YouGov poll for the Evening Standard has topline figures – with changes from last week’s poll – of JOHNSON 49%(+2), Livingstone 36%(-1), Paddick 10%(nc). Boris Johnson remains slightly shy of an overall majority, so taking into account second preference votes the figures become Johnson 56%, Livingstone 44% – so it appears this time the second preferences are breaking slightly in favour of Livingstone.

Following the fuss last week there is a slight difference in YouGov’s methodology to take into account ethnicity. YouGov are now weighting to 26% non-white Londoners. This is somewhat lower than the 29% ICM are aiming for. YouGov’s figures are based on the proportion of ethnic minorities over the age of 18 in the 2001 census. 29% is the proportion of ethnic minorities in London as a whole, since the non-white population is disproportionately young. I think it is also – by co-incidence – the GLA’s 2008 projection of the proportion of ethnic minorities of economically active age, which I guess is where ICM’s figure came from. Whoever is right, the slight difference between YouGov and ICM’s proportions of ethnic minorities is not going to make a hugely significant difference, perhaps a percentage point or two.

On the subject of the London elections, there’s a fun tool here that matches candidates to your policy preferences.


33 Responses to “Boris’s lead grows to 13 points”

  1. What figure has been used in the overall age weighting for over 55’s in this poll ?

  2. It says on Mike’s site – 28%.

  3. Does it ? I think it says they should be weighted to 28% not that they have been .

  4. The quote from Peter is that the remaining polls will have it weighted to 28%. But anyway, the weighting figures are here, and it was.

  5. Anthony, someone has posted on PB.Com that Labour are planning to report YouGov to Market Research Society of Great Britain. Whats your view on this development?

  6. That a dog returns to its vomit.

  7. :D

    If they do report YouGov, would it be an unprecedented development? I can’t remember a political party reporting a pollster for bias before, but my memory isn’t that good on these things.

  8. Thanks , Anthony , just a small quibble , the table does not show the voting intentions of the roughly 20% of voter’s who did not have a Party ID of the 3 main parties .
    I suspect that the poll I completed yesterday is not this one .

  9. Polls that show the lead to large would surely be bad for one’s chosen party surely?

    I remember the 1983 election as a kid and the Tory vote was actually a shade disappointing – despite the awful setback for Labour.
    I’m sure it had something to do with polls giving them high 40s and above.

    Surely a party wants a narrow lead showing it can win, but to keep supporters on their toes.

  10. Mark – they never do, which I think is rather annoying, but the splits are the standard ones that YouGov always produce. I would have liked to see the break by ethnic minority as well, but again – not there.

    It you did one yesterday then it certainly isn’t this one – the fieldwork closed on the 4th.

  11. The one I did had some very strange supplementary questions and asked me who I would vote for as London Mayor even though I do not live/vote in London

  12. If I was advising Labour I’d steer clear of reporting anyone.

    This whole thing is coming across increasingly like shooting the messenger and giving the appearance of the kind of news management and spin that Brown was supposed to put an end to.

    I can understand them wanting to counter bad news but this looks like another example of Labour being to focused on short term daily news management (all be it in an election campaign) rather than on the bigger issue which is there credibility with the electorate.

    Someone should perhaps take some of the campaign people aside and point out that if ,come May, these polls come true then they will look like they have tried to suppress the truth and bully polsters and that isn’t a good image to create in the last years of a government when you are behind in the polls.

    In the past I’ve had the same argument in the SNP when we have been faced with a bad poll and our press people have tried to rubbish the poll or polster.

    My advice is always to focus our message and a response to the bad news rather than to deny the news is bad.

    Peter.

  13. The Livingstone press release is here. The complaint seems to be on the grounds that the poll contains too many over 55s, which the last one did but this one doesn’t, and that it didn’t take account of the proportion of ethnic minorities in London, which it did.

    Bless his cotton socks, it’s so utterly without merit in relation to this poll that it only really evokes pity.

  14. Well it looks like Ken’s not taking my advice.

    Peter.

  15. Thanks for the VoteMatch Anthony. I thought it was really fun. I got Livingstone which is not really suprising but I diagreed with all the points about giving him more power which is perhaps also not suprising. Anyone else had a go yet? Does anyone know if there’s a similar site for the general election candidates?

  16. Could you explain the age weighting thing in more detail – I couldn’t follow the example below. As far as I can see YouGov’s panel has 50% of respondents over 55, and then they weight this down, wrongly originally, and now correctly to 28%. So it’s almost half.

    But this doesn’t make much difference to the result becauase of the party ID weighting. That’s the bit I don’t get, as it seems to me the unweighted sample has pretty similar party ID to the weighted sample.

  17. Well I managed to match most with UKIP which surprised me because I’ve nothing in common with them at all. After that it was the Libdems which is less surprising. I was closer to Ken than Boris, ( Oh….. sorry Steve I meant Mr Johnstone).

    Peter

  18. It wouldn’t be like this in an Independent Scotland… All figures would be up to date and graphed as if by magic…..

    Peter

  19. Actually following on from my earlier comment, I think I just don’t understand Yougov’s sampling method. As far as I can tell in every poll until now their unweighted sample had around 50% over 55s. Yet in the lastest poll their unweighted sample suddenly has fallen to 38%.

    Why would this change so much?

  20. Matt – as an aside, the raw sample demographics don’t correspondend to the panel demographics – it’s an assumption lots of people make but it’s not actually true. Raw samples aren’t just a random slice of the panel, so aren’t a microcosm of it.

    Weightings all affect one another – so when over 55s are weighted downwards, it reduces the proportion of the sample who are Conservative (and has effects on things like class, region and so on I expect), therefore Conservative identifiers need to be weighted back up so you can have a sample that contains both the correct proportions in terms of age, and the correct proportions in terms of party ID (and the correct proportions for all the other things).

    In practice the weighting programme will go round and round a number of iterations, adjusting the weighting figures until finally you get weights for each respondent that total up to the correct target weights for each demographic.

  21. In a perfect world there would be minimal difference between the weighted and unweighted figures. As I mentioned, YouGov’s samples aren’t just random people from the panel, they are generated to be as representative as possible to start with.

    In reality they still need weighting because of differential response rates – if you have the correct age proportions to start with, but only 33% of your under 55s respond and 66% of your over 55s do, you are going to have to weight down the over 55s to get back to a representative sample.

    If the raw proportion has fallen, it might mean YouGov decided their raw sample had too many old people and invited fewer, cutting down on all that wasteful weighting downwards, or it could just be co-incidence and they happened to get a very low response rate amongst over 55s last week. I don’t have anything to do with sampling, so I’m afraid I don’t actually know for sure.

  22. IF they polled on a Wednesday all the old codgers would be at B&Q getting their 10% discount.

    Peter.

  23. 1)Ken Livingstone has a real problem with the Evening Standard and probably wouldn’t believe its contents if he wrote them himself

    2)VoteMatch was fun – came out as UKIP followed by Boris with the Lib Dems and Ken scrapping for bottom place

    3)Surely the more raw data needs adjusting to fit the more likelihood there is for error?

    4) Ken is stupid to keep attacking the pollsters rather than getting his message across (although it seems to be largely about how much he doesn’t like boris)

  24. I don’t think the Tories will win by 13 points,
    but can anyone advise me what does happen if someone
    wins 50% on round 1, whoever it is?

  25. 50% on round 1, doesnt that make them Major straight away without taking into account second pref. After all they will have over 50% of the popular vote

  26. I don’t believe that Livingstone is 13% behind Johnson – but he definately WILL LOSE – my prediction is a lot closer .

    Dear Ken forgets how much his beloved Labour party has alienated his Muslim constituents , so he is foolish to rely on their vote to the maximum – they will reward him justly i am sure !!

  27. Can’t think of anything that the over 50’s have to thank him for either !

  28. About 2 weeks after the last GLA elections in 2004, stats were produced for each and every ward across London indicating how many votes per party were gained in each.

    Does anybody know where that resource is available?

  29. A bad night for Boris and Paddick on the Newsnight debate I thought. I wonder if it will effect their standing in the polls?

  30. Ken was more reasonable and polished, and was able to answer in a fairly interesting way. But he was basically repeating yesterday’s lines.

    Boris should have had his figures more at hand, but he is quite right to insist on a new Routemaster bus.

    Paxman was unfair in his treatment of Boris – and I’m not one of these Tories who moans about the BBC – read my posts on Con Home. He introduced him with a nasty little smear and was rudest to him in questions.

    I shouldn’t think it’ll make much impact one way or the other because Newsnight is probably watched by those of fixed views.

  31. Perhaps Paxman was rudest to Johnson because he was the one who waffled and postured most, and refused (or was incapable) of giving clear answers to many of the questions. This should give London voters something to think about – but I doubt that it will be reported accurately in the Standard!

  32. That’s not entirely correct – Paxman introduced him with a smear, whereas Paddick was merely asked politely how long he’d been a member of his party.

  33. There was an edge to that question as Paddick has not been a party member for long – but this was not followed up.