Over the last couple of weeks we’ve seen some apparently contrasting polls – have the Conservative lead dropped to single figures or not? Different polls have told different stories (and sometimes lead to people jumping the gun and thinking the Labour recovery has stalled or reversed). What is the actual picture?

Well, looking at polls since the bank bailout was announced in October the pollsters are actually all showing very similar levels of support for the Conservatives. Leaving aside BPIX whose high Conservative scores we can’t really speculate about given the lack of details about their methodology, other pollsters have ben pretty solid in showing the Tories in the 41%-43% range.

The difference is more in the Labour level of support, and here I think we are seeing a knock on effect from the Liberal Democrat and “Other” scores. The lower levels of Labour support have come from ICM and ComRes who are respectively the pollsters who show a higher level of Lib Dem support and a higher level of support for “Others”. YouGov, who tend to show the lowest level of Lib Dem support, have shown some of the narrowest leads for the Conservatives.

The wider pattern seems to be that the Conservative support, while dropping from their highest scores, has been quite resilient and stayed above 40%. The narrowing lead has largely come from non-Conservative supporters rallying to the Labour party. With that in mind it makes sense that the pollsters with the methodologies that show the highest support for Lib Dems and others will produce a lower level of Labour support.

40 Responses to “What is the real Tory lead?”

  1. I know you hate the rolling average but can you tell us what it is now after two new polls, I suspect it will be unchanged.

  2. Anthony, Re your above comments, I wonder why the Comres poll shows such a low poll for the Libs then this time?

    When an organistaion commissions a poll, can it have any affect on how margins of error are reported? In other words can an organisation commissioning a poll tell a pollster to reflect a margin of error in any particular way or affect any weighting in a chosen way or is that a big No-No?

  3. ComRes don’t normally show a particularly high Lib Dem score – they show a high “other” score. (Collectively others are actually above the Lib Dems in this poll – 13% to 12%. It’s made up of 3% SNP, 3% Green, 1% Plaid, 1% BNP, 1% UKIP, 4% “other others”)

    Gary – look on the front page of the site, it’ll always be there on the top right. I’m not going to put it up in posts because if you are here reading the detailed posts comparing individual polls with ones from the same pollster there is nothing more an average can tell you ;)

  4. Not sure what you mean about margin of error. The margin of error is based purely on the sample size – for a poll of 1000 people it is around about 3%.

    How they are reported is a different matter – newspapers don’t have to report them at all, indeed, I think they quite often aren’t reported in British newspapers.

  5. Anthony . Would you agree that at those times when there are major events influencing voter behaviour (mainly floating voters) what is critical is firstly precisely when the polls take place and secondly how quickly responders change their views because of the new events.? The floating voters represent a relatively small part of the total population so therefore the margin of error is higher than for the full population. Therefore overall greater variations between polls occur at times of a material change in sentiment regarding the political parties.

  6. Thanks, what I was wondering is if say a polling organisation produces a set of results which hypothetically shows 40:30:15 for 3 main parties and the pollster advises that there is a margin of error of 3%, can the organisation commissioning the poll ask the polling organisation to reflect that possible 3% error in a particular way i.e to produce a result of 40:30:12 for instance or to align any weighting in a particular way so as it benefits or detracts from a particular party or is there absolutely no influence allowed over the final figures at all by the commissioning organisation?

  7. Mike – I don’t think that follows. If there is a wider margin of error on a subgroup of people (and there is, of course) then that group also makes up a smaller part of the whole and any variation makes a smaller difference to the total figures.

    Richard – no, definitely not, absolutely no influence at all (it’s also worth pointing out that a 3% margin of error doesn’t mean that any figure within that +/- 3 margin is equally likely – the mid point, the point that is given in the figures – is the one that is most likely)

  8. Anthoney, is it not also true that those pollsters showing a smaller gap between Labour and the Conservatives also tend to show a drop in the Conservative figure (rather than just a drop in Lib Dems and Others)? Not that we can assume that these Conservatives have moved to Labour. And does Others include undecided? I was just wondering how big the pot of voters outside the main the main 3 is, so could we have a rock solid Conservative vaote of above 40% erroded by these undecided going into the Labour and Lib dem pots and thus reducing the Conservate lead without reducing the number of those voting Conservative? i.e. 40% Conservative, 35% Labour 16% Lib Dems but then we have an influx of undecideds into Labour and Lib Dems thus causing their % to rise and Conservative % to fall even though the NUMBER of Conservative voters has not changed.

  9. Gary, everyone has shown a drop in the Conservative share, but (putting aside BPIX) there doesn’t seem to be a significant difference between the Conservative share in the polls with big leads and the polls with little leads.

    Others doesn’t include “don’t knows” – they are excluded or reallocated (the “shy Tories” article under the FAQs has the run down) – so yes, theoretically that could be the case – a party could fall without losing any vote if “don’t knows” or unlikely to voters move to Labour or the Lib Dems.

  10. When a poll includes priming questions like “Are Cameron’s team lightweights” any statistics that eventuate are false readings.
    Any con-man, shuckster, psychic, magician and fortune teller knows how to use such questions to force a result.

  11. Victor – nope. Polls, including this one, always ask the voting intention questions first so that any subsequent questions cannot skew the results.

  12. Anthony. Surely Victor means that the answer to that particular question can be artificially skewed, not the answer to the voting intention question?I would think he is correct, if that is what he meant.

  13. I do think it’s best to offer some scale like


    rather than “do you think x is leightweight?”
    “do you think Labour/Tories (never LDs) are sleazy?”
    as the latter examples can rather plant an impression, even if, of course asked after the core voting questions

  14. It would be a fair assumption that we don’t need to ask Victor what his voting preference is.

  15. Collin – that’s a different matter, but Victor said “priming questions” so I thought it would be worth making clear that all these questions can’t affect voting intention.

    JJB – the scale doesn’t help much, what makes questions like useful is if they are asked about all the parties in the same fashion – then even a skewed question gives you comparative data.

  16. Anthony, comparative data may be useful, but not all comparative data is informative – I think JJB makes this point sufficiently well.

    If a skewed question is asked of all parties then each question about each party needs to be asked simultaneously, otherwise after the first repitition the questioner becomes primed and predictive, thus the value of the response changes. If successive questions are to retain value then they ought to be neutral.

    Surely this is similar to the order bias in which voters tend to favour those at the top of the ballot paper.

  17. Thomas, pollsters often rotate the order of parties / party leaders to mitigate the bias you identify. However, the purpose of a lot of the one-off political questions is not to get a precise measure of public opinion but instead are an attempt to guage reactions to a particular political point that a party may be making. These questions are often not symmetrical between parties — e.g. there’s not much point in asking “Do you think the trade unions have too much influence over the Conservatives?”. It is also clear that pollsters have different standards. Some pollsters are scrupulous in trying to balance the number of positive and negative questions, whilst others seem quite happy to ask a series of blatantly loaded questions on behalf of their clients.

  18. Not sure I’m convinced about that – with respect – don’t some people sense a certain amount of peer pressure to just say politicians are sleazy and lightweight because you’re putting it to them.
    The number of people in the pub who will come back and say no I don’t think politicians are sleazy and this is why is probably a minority.

  19. Thomas makes a good point on that, but I guess you can rotate the order of the parties.

  20. Rotating the order is lists is common practice, though if it’s a question where people have strong feelings it doesn’t actually make much difference. Where it makes more difference if big long lists, especially ones where people can tick lots of things.

    Incidentally, British studies have found an alphabetic bias only *within* party slates – in a multimember ward people at the start of the alphabet will get more votes than candidates from the same party with names later in the alphabet. In single member wards being at the top of the ballot paper confers no advantage.

  21. Bah – I got more of the split votes when I stood for election – although all three of us were heavily beaten. I wanted to believe it was some reward for my efforts, now I’m told I was just top of the ballot paper.
    BTW, Anthony, didn’t you win on the split votes in Dartford, but you must be W?

  22. Sorry – delete that post if you want, it was unfair of me to raise your own election.

  23. Obviously there are other factors that affect it – foriegn names, addresses outside the ward, name recognition, incumbency – presumably canvassing and hard work.

    I did indeed come top of my slate, but modesty prevents me speculating on why!

  24. I feel a bit guilty about one of the other candidates who got about 150 votes less, totally unfairly, because he canvassed quite a lot although think he was really a bit paper. But he had a foreign name and was buried below quite a few other single candidates that generated quite a lot of split votes.

  25. Weighted Moving Average is 42:33:15. Whether people will still be buying the “global leadership” line when it is obvious that the UK is in a more serious recession than the rest of the G8, despite the near-collapse of sterling, remains to be seen….

  26. I polled bottom of my slate in 1998 when Labour won Mortlake Ward, Richmond-upon-Thames (just). I am above (Brian) Matthews alphabetically but clearly my surname doesn’t sound very English : my cousin Barry Langford is every bit as Jewish as me but because he has a totally Anglicised name and is ahead alphabetically he got more votes than I did. Brian’s known tenure of a Fulham season ticket probably helped too! So, Barry & Brian won & I didn’t.

  27. I would say these polls are pretty good news for the Tories, although it doesn’t feel that way in many repsects. To retain scores of 40+ must be a big positive against a backdrop of some less than gushing media coverage, but seeing Labour recover with relentlessly bad economic news is a worry for Cameron.

    One thing that does strike me, and perhaps the right leaning posters here can comment, is my impression that the Tory party is remarkably twitchy. Over the last 18 months we’ve seen on several occasions if not signs of outright panic then at least disquiet when polls run slightly the wrong way, even if the overall position is good. Labour by contrast had to plumb the real depths before they started to really twitch, but that is now largely put behind them, despite being clearly second place and heading for a defeat.

    I wonder how this will translate if and when Cameron forms the next government. The Tories biggest asset down the years was a relentless ability to pull together and not panic at difficult times. It looks to me that they have lost that knack, and gaining power with a small majority might prove very difficult for them.

  28. I agree some in the Tories are twitchy.

    One only has to observe Con home (if many are real active Tories).

    As soon as a slightly bad poll appears the knives are out.

    Perhaps suddenly going from an inadequate lead to a large lead in 2008 with little period in between made it seem unreal.

  29. I sometimes compare party leaders to the characters in ‘Dr Finlay’s Casebook’ (more the TV series than the books).

    Is it possible that in sunnier economic times people wanted young Dr Finlay (Blair, Cameron, Clegg) but in these harsher days people prefer the reassuring figure of old Dr Cameron (Brown, Clegg, who is the Tory (Clarke?)).

  30. The second Clegg should have been Cable.

  31. Pinch of salt times – certainly nothing the Tories should be worried about – any perceived Labour gain is all at the expense of the Liberals.

  32. According to YouGov 5% of the vote is going to either the BNP or UKIP which given that these voters may be assumed to be more right wing than left wing should be of some concern to the Tories. Is that also to taken with a pinch of salt MIke?

  33. Shouldn’t the question be what is the real LibDem vote?

    It seems that the LibDem vote is more unsteady than any of the others. In Glenrothes the fall in LibDem votes accounted for the outcome, but why are they doing so badly?

    The LibDems are ignored except Vince (I told you so) Cable because their leader is not going to be the person who diplaces the incumbent PM (or not) .

    Is that all it is? Will their support in the election turn out to be not much below the last election especially in the seats they have held for more than one parliament?

  34. Would it be possible to simply assume that tribalism is in decline and that all parties have a weaker core support?

    It would however be interesting to know how the tribalist vote stacks up proportionally.

  35. Anthony,

    Sorry to be a pest but the “Detailed Results” link isn’t working again.

    Maybe you should redesign YouGovs web site as well.


  36. Mike Richardson – I would disagree that the Tories shouldn’tbe worried. The current numbers are still good, but underlying perceptions are weakening, and a key issue is the fact that they have been wrong footed and seen as not having a clear philosophy. This will hurt them if they can’t put it right. Take 2 – 3% off their current score and they are in trouble.

  37. A glance at the regional breakdown for the two polls showing the narrowest Tory leads (YouGov and Populus) makes it doubly hard to work out what is going on out there. Granted the sub samples are small but for Populus to suggest that the Tory rating in Scotland is at 9pc the lowest its been all year and for YouGov to suggest that on the contrary it is at 21% the highest it has been for simply ages is bizarre. The English regions are no less confused.
    I think we are going through a period when the polls cannot measure the mood far less the voting intentions of the public and only in the New Year will they settle down. Should the Tories be worried? Certainly they should be concerned that the government has dominated the media for the last two months but although Alec above says the underlying perceptions behind their numbers are weakening the underlying reasons for the governments unpopularity have not gone away and will sooner or later resurface. How quickly they do so may indeed depend on the Tories ability to establish their take firmly in the eyes of their would be supporters. The stand against borrowing our way out of the recession may succeed in doing just that since it establishes clear blue water between the two main parties. Time will tell.

  38. I think the Tories are taking the right line actually, much as I’d love to see bigger tax cuts.

    Tax cuts do bring in more revenue – and are essential, but the trouble with just spending the same and cutting taxes in this situation is you will send the markets into a fright that the government is even more in debt.

    Janet Daley was arguing we should just slash them because they always bring in more revenue. It’s not serious politics. If we had balanced the books better up to now, we would be able to, but the cupboard is bare and we are where we are.
    Margaret Thatcher knew that in 1979.

    I’d like to see some really urgent measures to get credit to businesses.

  39. There is an alternative.
    Cut spending severely.
    Freeze NHS, Education.
    cut defence, cancel Crossrail, cut benefits.
    But I don’t think anyone is suggesting we do.

  40. “I’d like to see some really urgent measures to get credit to businesses.”

    Yes,there is still a problem with that. The Banks have done a 180% turn on attitude to risk , which is overdue, but seem to be applying their new found prudence with great inflexibility.

    I have severe doubts as to the actual impact of tax rebates on consumer spending, and through to jobs.

    How much of the mooted tax refunds will be used to reduce debt, rather than being spent?

    Which sectors of the economy will this money be spent in?
    Not housing construction-since that needs mortgage availability & 20% deposits now.
    Not Financial Services since it is this sector which is contracting from pre-boom levels.
    Consumer Electronics? -mostly imported.
    Food & Clothing?-people are shifting to Lidl & Primark in droves to reduce household costs-so where will jobs be saved in these sectors?

    I hear the bold assertions from Mandelson & Brown linking “Tax Reductions” with “Getting the economy moving”. I have yet to hear how either of them think this will work.

    Brown mentions USA & China on Fiscal Stimulus as though they aid his case.

    The $170 billion package in USA has had no discernible effect on there worsening economy that I can see reported.Now they plan to bail out the US motor industry, whose labour costs are internationally uncompetitive .Yet Brown-cheerfully ignoring the British Leyland fiasco-supports this proposal!

    China’s stimulus package of $600 billion is being funded from Reserves of £2 Trillion-not borrowing.An interesting FT analysis questions it’s effects on global economies-and of course points out that China’s Banks are owned by The State!

    All the same-will UK voters willingly grab their tax rebate before Christmas….and then increase their support for the Conservatives in the Polls?

    Brown’s answer to that question must be “No”-and who could believe him to be wrong?