The Evening Standard has London voting intention figures based on the aggregated data from the YouGov daily polling over the last week or so (I should add that the raw data is aggregated as if it were all one poll, and then collectively weighted to London targets – so it is the equivalent of doing a London poll, just with the sample invited in lots of little chunks rather than one big go).

The topline figures are CON 39%, LAB 35%, LDEM 17%. I’m not sure what figures Joe Murphy is thinking of when he says the lead has halved in six months – at first glance I can’t find any London polling from last summer/autumn – but a YouGov poll in April 2009 showed a Conservative lead of 12 points, so it has dropped by 8 points since then. YouGov national polls were showing a Conservative lead of 18 points at the time, so if anything thing have narrowed slightly less in London than elsewhere.

The four point lead represents a swing of 5.5%, so slightly more than the swing suggested by YouGov’s national polls. On a uniform swing the Conservatives would gain 8 seats from Labour and 3 seats from the Lib Dems (depending on the notional figures you use – on Rallings and Thrasher Poplar and Limehouse would also go).


84 Responses to “YouGov voting intention for London”

1 2
  1. NEILA

    “The difference between polling companies is irrelevant. Its the differential between national and marginal polling by an individual company that matters.”

    That is what I meant !!

    AR marginal poll 12% swing/ National poll 7.5% swing = differential of 4.5%

    Expect the same kind of differential if the ‘landlside via the marginals’ id a respectable positon.

    YG swing last night was 4%- so the YG poll of the marginals on those numbers needs to be at least 8.5% for this story to have real legs.

    Given the AR-YG differential I would expect it to be even less than 8.5% when it comes and the London poll today – a key battleground with many marginals- shows a swing of 5.5% and supports the cautious perspective on Conservative marginal performance.

  2. I you are tired of Richard O’ PPB then you can always tune to Dutch TV if you can get it and follow local elections results coming in on NOS!

    QI

  3. @Maurice any minute

  4. CON 38% (38)
    LAB 32% (33)
    LD 19%(16)

    Apart from the weekend blip, the Con lead has been stable at approx 6% for the past 7-10 days.

    No apparent impact from Ashcroft (unsurprising as the hypocrisy was unbelievable)

    Has’ labours reurgence stalled?

    What will happen when the Conservatives start campaigning?

  5. 38-32-19: one out on LD !

  6. Sorry Rob, your post wasn’t clear.

    I think you have to look at the differential in proportional terms though. By which I mean you would expect the difference between marginal and national polling to narrow as the overall swing narrows.

    If there was a 1% swing in national polling but a 2% swing in marginals that would potentially be more of a guide than a 5% swing in national polling and a 9% swing in marginals.

  7. Suprised by the 3% increase for the LibDems – wondered what’s caused that?

    Oh well, tomorrow night will reveal all I guess…

  8. 38-32-19 gives a UNS projection

    Cons 287
    Lab 285
    Lib 48

    No real effect from Ashcroft story then.

  9. @Rob Sheffield
    Is that tonights result?

  10. Much the same. I think Anthony’s right about the 6% lead being fairly stable at the moment. And the LD score is more plausible than one or two previous showings. Not much of a point to be made, perhaps, but does it slightly support the notion that if LD support rises it might be at the expense of Labour rather than the Tories?

  11. 7% 5% 6% Tory Leads in the last 3 days.

    Need to see a few more days to determine if the trend has stalled.- or if it will continue at a slower pace.

  12. @TONYM

    “No apparent impact from Ashcroft”

    As myself and others- with different poltical views like Roland- have posted there just is not any currency in this personal destruction BS- did not work a week ago with Brown and its not working with Ashcroft now (and won’t).

    As you said the 6% +/- MOE lead is becomig established with YG now plus tonight a more accurate LD number IMHO.

    But Tories not yet started campaigning….oh come off it !!!

  13. Richard, have you ever considered why policemen retire at 50? Could it perhaps be because after that they rapidly stop becoming fit enough to chase down and detain twentysomething criminals?

    I don’t really think Thatcher would back you up here – her big idea in 1979 was pay rises for the police.

    I’d be interested in the regional breakdown here. I could see a big differential broadly conforming to an inner/outer London split, but probably also with Labour beating the average slightly in old heartland bits of East London and falling short throughout inner West London.

  14. LDs back to 19.

    Swings of 3 from 16 to 19 to 16 and back to 19 again are quite big from such a low base.

    Yet we are expected to think these polls are sensible.

    ‘Others’ have moved from 13 to 11 and its all gone to the LDs?? UKIP and BNP gone to the LDs?

    Yes yes – I know, I am just dim and all these apologists for polls have brains the size of planets.

  15. Actually I’ve been punching numbers into the swingometer and an increase in LD support costs the Conservatives more seats than Labour even if the two main parties stay on the same percentages.

    This must be because there are more Tory.LD marginals than Labour/LD marginals.

  16. @trevorsden
    Thanks for your support Mr T, I really dont know why he thinks I’ve misquoted him.

  17. Sheffield – all the stories about Brown came from a left wing journalist and a left wing newspaper. Nothing to do with conservative campaigning.

  18. @NEILA

    “but does it slightly support the notion that if LD support rises it might be at the expense of Labour rather than the Tories?”

    It will do in polls- John Curtice projects roughly 2-1 preference for Lab of LD’s as second party so you would expect a rising poll lLD poll number will impact lab more.

    But it also means Tactical Vote of LD’s at GE will benefit Labour by same proportion roughly.

  19. @Tony M ‘No apparent impact from Ashcroft (unsurprising as the hypocrisy was unbelievable)’

    On the contrary, it was only too believable!

    Anyway, we know that you can rarely read any single poll as a referendum on any specific issue. The issues go into the mix, and some stick, others have legs, and most die the death.

    I think the Ashcroft affair will detract a little more from the electors’ view of the Tories’ judgement, reliability and frankness, and will hinder their recovery in the polls.

  20. Veering completely off topic, as the police have been mentioned and I am one…..

    Policemen used to retire at 30 years service, now it is 35. The change is not retrospective (I joined at 21 so I retire at 51). Edward is right that after 30 years only a minority of officers would be physically fit enough to discharge the duties of a constable, although many retire and are rehired as civilians (retaining their skills in the force).

    The big pay rises the police enjoyed under Thatcher were actually nothing to do with her. The Edmund-Davies formula, which increased our pay over several years, was brought in in 1978 under Labour. The miner’s strike brought huge overtime payments to many officers and I think people get the two issues (salary increase = Labour, massive net pay rate = Thatcher) confused.

  21. Your right Tony M, the hypocrisy is disgusting.
    The same as the Country needed a change after the tired last Conservative Government in 1997, we need a change now. If it doesn’t work out, then people can always vote Labour back in again in 5 years time.
    Back to the hypocrisy;
    It’s in several papers today that Lord Paul, a major backer of Labour, and also a non-dom, bought 6000 copies of Gordon Brown’s ironically titled book ‘Courage’, before being made a Privy Counsellor.
    Clearly the non-dom issue is therefore one for all parties, not just the Conservatives.
    Courage eh, you couldn’t make it up…….

  22. Persoj called trevor

    “Sheffield – all the stories about Brown came from a left wing journalist and a left wing newspaper. Nothing to do with conservative campaigning.”

    what point are you on earth trying to make- you sound a bit befuddled.

    The conservatives have started campaigning for the election was the point that I was making as Tony M said they had not yet !!

  23. @TonyM
    2What will happen when the Conservatives start campaigning?”

    Recent history would suggest not much.The party in the lead at the start of an election campaign usually sees their lead reduced up until polling day.The nightmare for the Tories is if the 1992 effect returns, an electorate unsure about the alternative flooding to the government as a last minute bout of nerves kick in.

    In my opinion it’s unlikely Cameron can make enough/any progress from this base to get a majority.

  24. The earlier poll for London was 46-33-16. — totalling 95, leaving 5 for others.

    The latest one totals 91. 11 others.

    Are we to believe that posts expenses the toiries were 46 ?

    When all parties especially tory/lab went down significantly nationally ?

    The change from 2005 is significant enough.

  25. geriant

    I think your wrong on that as 70% of economists say that we need to cut now or we face a double dip reccesion. Most do not back brown on his economic policy and I think April 23rd will prove it anyway as the new figures will show we are back in reccesion.

  26. Amber

    What are your thoughts? You normally give a full analysis. Do you think the trend has stalled? Or do you think it will flat line and continue or reverse?

  27. Anthony, you say that the Conservative lead has narrowed slightly less in London than elsewhere.

    A 12% lead to a 4% lead is a fall of 66%. An 18% lead to a 6% lead is a fall of 66%.

    If anything the figures mirror the national figures.

  28. This London poll being an aggregate of the London subsamples from national polls is completely worthless .
    The last 6 Yougov polls have had London subsamples of
    Con/Lab/LD
    43-32-14
    29-39-22
    39-30-24
    47-29-14
    39-30-18
    36-35-22

    Con range 29-47 Lab range 29-39 LD range 14-24

    No scientist or psephologist would aggregate such disparate figures and claim to have a meaningful estimate of London voting intentions .
    FWIW these figures average out at Con 38.9 Lab 32.5 LD 19.0
    THe above just illustrates how worthless looking at subsamples actually is and aggregating worthless figures still gives a worthless result .

  29. Mark, that’s nonsense.

    Subsamples are worthless purely because they are small and because they are not properly weighted. If you amalgamate them together and weight them properly it solves these problems.

    Every meaningful sample of 1000 contains within it lots of little sample of 10, which taken alone would be meaningless – so of course something meaningful can be built out of things which taken alone would be worthless. If you took a properly weighted sample of 1000 people, then dismembered it into 10 little samples, and took off the weighting they would look as disparate as those subsamples. It would not invalidate the original sample.

  30. CLAD – wrong way to look at it. Even if you looked proportionally, you’d want to look at the shares of the vote, not the difference between them (though swing doesn’t tend to be proportionate anyway)

  31. Not much encouragement for the SNP here.

  32. I take your point Anthony, but you only mentioned the leads, in your op-ed, not the share of the votes.

  33. CLAD – because I didn’t do it proportionally ;)

  34. Sorry Anthony but you are competely wrong , take out 1 of the 6 subsamples at random would still leave you with an aggregate 900 or so sample which according to you would be a pretty meaningful total sample size . The figures would vary between
    Con 37.2 Lab 33.4 LD 20 and Con 40.8 Lab 31.2 LD 18.4 and points in between .
    Take out one poll at one extreme and replace with a poll at the other extreme and the variation would be much larger a bit too late for me to work out sample figures .
    The example you give of dividing a bigl poll into small subsamples is a red herring .

1 2