There is a new ComRes poll in the Independent tomorrow with topline figures – with changes from ComRes’s most recent poll in the Sunday Indy – of CON 39%(-1), LAB 31%(nc), LDEM 16%(nc). The poll was conducted between the 24th and 26th October.

Clearly there is no significant change from the this previous poll, but that in itself is significant as since the last lot of polls the media narrative has moved away from blanket coverage of the banking crisis to more normal politics (in this case, George Osborne, Peter Mandleson and a rich Russian’s yacht…)

More tomorrow…


54 Responses to “ComRes show 8 point Tory lead”

1 2
  1. No, I mentioned hoarders, and I don’t like the idea of hording money – any more than I like the Bullingdon alumni and their criminal activities – no doubt some went on to do some good, not many though. A by-word for excess, disrespect and thoughtlessness. Is it still going?

    A mistake in teenage years ought not to halt anyone’s progress completely. Unfortunately, a privileged background lets you off the hook a lot more easily than a poverty-stricken one. I find it all disgusting, and care little who thinks I’m irrational.

    I wonder what happened to the kids who were expelled for the drug-taking.

  2. As a preliminary aside, this poll appeared in a particularly excellent edition of “The Independent”, excellent in my view because it also contained a letter from me about the inadequate capacity of the Channel Tunnel Rail Link through Kent, particularly between Stratford and Ebbsfleet. Well, maybe that’s not such an aside because of the concentration of marginal seats in North Kent where transport problems may swing votes. If the result if as close as this poll suggests, North Kent issues could even decide the result. The figures in this poll put, for instance, the Medway Towns seats very much back in play. (Joe James B’s earlier comment, amongst others, is relevant to this point).

    However, let’s look at the psephology proper of this poll: there are several things to think about.

    Firstly, the poll suggests that the swing to Labour is particularly pronounced amongst men and amongst AB voters. If we remember Anthony’s excellent poll of marginals earlier this Autumn, it suggested that the ordinary polls were underpredicting the number of seat the Conservatives would win because the reasonably, but not very, safe Labour seats tend to have more skilled working class electors, who appeared to be deserting Labour. If this poll is right in suggesting that the voters Labour is reclaiming are male ABs, I think one might extrapolate from Anthony’s data to guess that the voters who have change allegiance are disproportionately in safe Tory seats and/or in Tory-LibDem marginals where anti-Tory voters are less likely to vote LibDem tactically in a crisis. If this is so, a crude translation of the opinion poll percentages into predicted seats won is likely to underestimate the likely number of Tory gains at the next General Election.

    Secondly, the newpaper article accompanying these figures consider in detail how the economic crisis is changing votes. But conventionally, opinion poll data is analysed demographically, by age, sex and social class. Maybe we ought to analyse opinion poll data by concurrently collected economic data such as the TSE indices, unemployment figures and interest rates (as well as by demographics).

    Thridly, if we want to identify how the economic crisis is affecting voting intentions, we need to analyse lag effects statistically. I thought one reason why Anthony’s poll of marginals suggested a heavy working class shift against Labour was that middle class voters had already gone before 1005, e.g. over Iraq. Perhaps male ABs assimilate political developments intellectually, e.g. through the media, whilst women and CDE’s react when they are direcly affected themselves, or their emotional climate changes. In other words, there may be differentially lagged interactions between economic and demographic factors affecting voting intention: if polls are going to be acurate predictors we need to tease these out.

    Fourthly, the polls reported in the papers factor analyse a data set including responses to voting intentions, demographic variables etc. Such analyses included a number of assumptions, such as that the relationships between variables are linear (or identifyable curvilinear so that we can do appropritate transformations). To be somewhat cynical, such techniques fit what we are able to do both in relation to the development of statistics as techniques and in relation to the skills and abilities of those who do the work and interpret it.
    Unfortunately, if we suggest that voting intentions are being shaped at present by economic factors, this is to suggest that the future voting is going to be determined by causes which are very obviously behaving chaotically at present, not only in everyday language but in relation to the ideas of mathematics. This means that to be useful predictors polls must use appropriate techniques for chaotic data. Whether such techniques are available is doubtful, which is perhaps why pollsters find increasing use for qualitative techniques.
    As a matter of informal observation, voting intentions in recent years do seem to do through periods of stability, followed by rapid change. thus the polls stayed largely static between withdrawal from the ERM in 1992 and the 1997 General Election. More recently, we have seen changes in the polls when Gordon Brown decided not to call an election, and just lately following the economic crisis. The implication is that whether or not current polls will accurately predict the next General Election depends upon whether there is stability in voting intentions before 2010, or whether there is one of the sudden changes that can theoretically be modelled using non-lienar mathematical models (but for which it unfortunately effectively impossible to turn actual data inot accurate predictions).

    Opinion polls are important at a time of economic crisis not only for their interest in relation to voting, but also for their indications of popular opinion on matters that will affect the UK’s ability to recover. We have for many years now had almost conventionally constructed and administered opinion polls. What I am suggesting in what I realise is a rather technical post is that if we are going to inform our views on psephological matters, and arguably other more important ones, we are going to have to look more deeeply at the adequacy of the polls we are using and at how far we can interpret them.

  3. Not wanting to appear dismissive – I have a short question. What do you think would happen to the “don’t knows” figures if the words “if there were a general election to-morrow” were excised, and a straight “how will you vote at the next election asked”?

    The problem with arguments about the latest polls is that one seeks to project forward beyond the hypothetical to-morrow, and use meaningless phrases like “ave it!” too much.

    I’m against the idea of using polls to direct policy. Of course the Govt wants to do the popular thing, but in my view popularity should come after the right decisions anyway, without the stymying effect of second-guesswork.

  4. Alasdair, my apologies about that lapse, you are correct. I think it my P.R. side coming out, very hard to think rationally. However you can and I will argue that the person who leads the party has a factor on how people will vote. The blair ‘superstar’ factor is a case in point, maggie thatcher V kinnock.
    My main point is that I think Gordon Brown will be a similar divisive force as MArgaret Thatcher. However I am seeing similarities betweeen Gordon Brown and the labour party now and the Tory Party in the late eighties. I will not be surprised if Gordon Brown wins the next election, (being a liberal man it makes me spit to sya it) but I think he will be humstrung by a very much reduced majority plus the effects of this recession, with him having raising taxes and having to reduce public exprenditure..sorry investment.

1 2