A YouGov poll in the Sunday times has voting intention, with changes from YouGov’s last poll, at CON 39%(nc),LAB 31%(-1), LDEM 16%(-1). The poll was conducted between Wednesday and Thursday, with the overwhelming majority of the fieldwork carried out after the news of the Iran hostage release. Full tables are available here.

A hypothetical question asking how people would vote with Brown as Labour lead has a slightly larger Tory lead, with the Conservatives on 40%, Labour on 30% and the Lib Dems on 14%. As ever, it is debatable how meaningful questions like this are as normal voting intentions don’t mention the party leaders: it could be just as much a Cameron or Campbell effect. In this case YouGov have asked a second hypothetical question asking how people would vote at the next election if Tony Blair remained Labour leader: CON 39%, LAB 32%, LDEM 15%. The question still can’t predict how Gordon Brown will actually score in the polls – people may see him differently once he is actually PM, but it does at least provide comparable figures – a Conservative lead of 10 points with Brown, or 7 points with Blair.

Interestingly, it tells a different story from when Populus did a similar experiment last year, when mentioning Blair and Cameron increased the Conservative lead compared to the standard question, but mentioning Brown increased it more. Now mentioning Blair and Cameron decreases the Conserative lead compared to the standard question. My guess is that the reason is that the ansers people give to pollsters when asked the standard voting intention is now starting to take into account the assumption that Brown will be Labour leader very soon, so is starting to converge towards the hypothetical Brown figures, with a Blair hypothetical question now showing a lower lead.

Following the recent budget and the negative press over the 1997 tax on pension funds, Brown also has a negative approval rating for his performance as Chancellor – 41% think Brown is doing a good well as Chancellor, 52% think he doing badly – a net approval rating of minus 11, the first time in a YouGov/Sunday Times poll that he has recorded a negative approval rating in his role as Chancellor. Tony Blair’s approval rating is minus 27, Cameron’s is plus 23 (his best rating since April last year) and Sir Menzies Campbell’s minus 18.

Asked who people would like to see succeeding Tony Blair as Labour leader, 22% of people perferred Gordon Brown, with David Miliband on 16%, Charles Clarke on 7%, Michael Meacher on 4% and John McDonnell on 2% – 49% of people said they didn’t know. David Miliband’s closeness to Brown in the question is largely down to opposition voters – amongst Labour supporters Brown is still the runaway favourite, on 55% compared to 12% for Miliband – though of course, with Labour 8 points behind in the headline voting intention figures many of those oppostion voters are people whose support they need to win back to secure a fourth term.

There is also an ICM poll in the News of the World and a BPIX poll (albeit on conducted in the middle of March) in the Observer which I’ll look at properly tomorrow.

17 Responses to “YouGov in the Sunday Times”

  1. I believe that 69% of those over 65 believe Brown is responsible for the pension shortfalls. That could be a very meaningful figure.

    The £5bn/year might have seemed easy money 10 years ago, but can any party win an election if the elderly (who’re far more likely to vote, and are increasing in numbers) turn against them?

    16% support for Milliband is also now approaching very close to Brown’s 22%. Brown’s greatest asset at the moment is the appearance of inevitability, but I think that like David Davis, if a challenger catches him in the polls then his support could evaporate.

  2. if you compound the pension tax grab,it becomes one of the biggest taxes of all time,approaching £150 billion.this figure will be used more in future.
    this is cronic mismanagement and only brown and balls are to blame.
    the only reason for the different response between young and old regarding their negative view of brown is, the old are drawing their pensions and can see the difference.
    of course if you work for the state you are laughing as they are not effected.

  3. I agree with Philip Thompsons remarks.

    But one thing perhaps in Browns favour is that he has dominated the Treasury over the last 10 years and his removal from that job – which will occur whether or not he succeeds Blair as PM – might send shock waves through International Financial markets.

  4. Brown gives an increased LABOUR lead?

  5. Oops – thanks Paul. That’s what I get for posting in the middle of the night.

  6. The table is showing 4% at Westminster for the SNP. but a massive 38% in Scotland. Even as a SNP member I can’t see us getting that kind of vote at the next general election. It’s obviously skewed quite considerably by the current scottish election campaign.

    God knows how many seats that would give us at Westminster if it was actually to happen, as it would have us 9 points ahead of Labour ( with the tories now ahead of the Libdems). It could certainly cost Gordon Brown his place as PM.

    Having said that a pol in todays scottish mail on Sunday has the SNP on 40% for holyrood, although it’s only a poll of 1,000 and there are no details on who conducted it or the methodology, so the verdict as with MRUK is still out.



  7. Any thoughts?
    I notice within the detailed data tables given from the yougov survey that the Welsh numbers seen rather a surprise ie. con 38%, lab 31, lib 12 !!!
    Surely this is grossly incorrect? Or does a weighted bias have to adjust this?

  8. Jerry,

    It also shows PC at only 8% which seems odd to me given other polls and there previous level of support.


  9. Peter, they only interviewed 192 respondents in Scotland ans 114 in Wales. This makes the margin of error truly immense.

    In addition, those Scottish and Welsh subsamples will not be properly weighed, even by elementary things like age group, social grade and gender, let alone past vote. I strongly recommend that people don’t put much confidence on sub-samples like this.

  10. As Stuart says, the regional subsamples within national polls are really not reliable enought to draw conclusions about voting intention from, so it’s really not worth trying to read too much into them.

  11. Anthony,

    All the more reason to do a “No Payment” YouGov 25,000 Scottish member mega poll the weekend before the Scottish elections.

    It won’t cost you much and as the biggest scottish poll ever it will probably get YouGov the kind of publicity they can’t afford to buy.

    A poll of that size could well have 300 plus in every Holyrood seat, ( If you asked a Westminster question, it would be 500 near in all 56) It would let you give accurate regional predictions and would probably pick up the likes of margo macDonald and Tommy sheridan.

    You know it makes sense.


  12. Does YouGov have enough members in its panel to get a representative 25k of Scots? Without running the risk that a poll of that size, and that limited, becomes self-selecting.

  13. Philip: nope. The “self-selection” element would be ridiculously high.

  14. The overall panel is about 145000 now, I can’t remember what percentage of the population live in Scotland (and, of course, the panel as a whole isn’t necessary proportional) but I doubt it’s high enough for there to be 25,000 YouGov panellists in Scotland.

  15. Anthony,

    I seem to recall the figure of 25,000 being mentioned in one of your previous post, thats why I used it, but I could be wrong.

    As to self selection, as YouGov has a profile of every member filtering it to reflect the general population shouldn’t be out of the question, after all they already use that system before selecting people for each individual poll.

    I don’t know how sophisticated the internal software is, but I suspect that from my own profile and previous responses, YouGov could probably predict my answers to a Scottish political poll to over 90% accuracy.

    As you have said before YouGov has a sophisticated set of tools to prevent swamping, multiple indentities or fraud. The same system that stops people skewing a poll should be able to create a balanced sample from a megapoll.

    Therefore balancing all poll returns by individual profile to get a balanced sample before looking at the actual answers shouldn’t be an unsurmountable problem.

    It’s yet another benefit of an internet based standing panel over tradional random phone polling.


  16. If its 145k total, then I’d say 25k of that total would be far too self-selected. A 1/6 chance of being selected is very high.

    25k of the Scottish proportion of the 145k would be incredibly self-selected. Indeed, unless there’s a disproportionate amount of Scots on the panel then it would be almost every Scot who has ever signed up.

    Maybe 4 years from now at the next election, I agree in theory it should be possible, but I don’t think it is this time around.

  17. Philip,

    You could be right, ifthe panel is only 145,000, then even with filtering the largest poll we would see would be in the region of 10,000 ( I don’t know what margin of error that would give).

    It would give us some good indications of regional trends, but with samples of 1,000 or so in each probably only to within 4% or 5%, which wouldn’t be enough. It certainly wouldn’t let us predict seats.

    What might be a more valid project would be a 100,000 person poll by the time of the next general election, with people selected on the basis of past YouGov responses. Although again that wouldn’t be near big enough to tell us about individual seats.