ComRes September Poll

There is a new ComRes poll for the Independent out tomorrow. The topline voting intentions, with changes from ComRes’s last poll which was done immediately after the Lib Dem conference, are CON 41%(+2), LAB 29%(+2), LDEM 18%(-3). We’ve got the Lib Dem conference boost fading, and something of a boost for the Labour party.

It’s clearly a smaller boost for Labour than the weekend polls displayed, but then the fieldwork was done a couple of days later when Gordon Brown’s conference address was no longer ringing in respondents’ ears. It’s also still a comparatively low Conservative lead as the Conservatives have not recovered the support they lost to the Lib Dems in ComRes’s last poll. Still, our next polls should be post the Conservative conference and we’ll be able to see what effect their conference has – assuming they aren’t completely pushed off the news agenda by the economic turmoil.

43 Responses to “ComRes September Poll”

  1. Comparing the 2 recent Comres polls , their weighting for past vote is baffling .
    The 1st poll had voter % recall of how they voted at the 2005 GE as Con 18% Lab 25% LD 14% . This compares with ICM weightings of 19-22-13 and Populus of 19-23-12 , so a slightly pro Labour and LibDem bias in the sample . Comres after weighting had Conservatives almost unchanged , Labour a little down both of which seem reasonable but an increase in LibDems after adjustment whereas you may have expected a downward correction because of the apparent LibDem bias in the sample .
    The latest poll has voter recall of Con 20% Lab 27% and LD only 10% so a sample with a pro Con and Labour bias and quite a large under sample of LibDems . After weighting the adjustments are a slight increase in Conservative figures ( where you would expect a small decrease ) a small decrease in Labour figures ( where you would expect a bigger decrease ) and a small increase in the LibDem figures ( where you would expect a larger increase ) .
    It looks to me that the first poll slightly overstated the LibDems in the published figure of 21% by around 1% but the 2nd poll understates the LibDems in the published figure by around 2% .
    ICM would have given the LibDems a higher published figure in both cases . Some clues can be garnered as to which of these polls is more likely to be correct by the voter changes from 2005
    1st poll LD gain 24 from Lab but lose 6 net + 18
    LD gain 4 from Con but lose 20 net – 16
    2nd poll LD gain 22 from Lab but lose 9 net + 13
    LD gain 8 from Con but lose 14 net – 6

    The 2nd poll figures are clearly better for the LibDems than the 1st poll despite the headline figure being 18% instead of 21% .

  2. Mark Senior in “better poll for Lib Dems is the more accurate” declaration shock.

  3. Its unfortunate for the tories that their conference is taking place when seismic events are happening in the financial world. This means they’ll get less press than normal and will probably lessen their conference bounce.

  4. Not again! Can we please not have stuff about Labour/Tory/Lib Dem “lies” here? There are plenty of sites for such comment, but surely not this one. It has no relevance to the polls, which is what we’re discussing/absorbing here.

  5. Mark Senior

    Mark if you don’t mind me saying so that was a most unwise post. You are simply setting yourself up for a free hit when you launch into yet more convoluted arguments about why the polls are actually better for the Lib Dems than what they seem or biased against them for one reason or another. So Com Res has now joined YouGov in your cast of villans ! Who will be next?

    If the Lib Dems are doing better than the polls suggest then this should show itself in the upcoming Glenrothes by election but if as I suspect they lose their deposit it will be doubly hard for anyone to argue that is the case.

  6. Sceptic, you need to read the post, it’s the other way around.
    Mark’s point is that the second Comres poll seems to be better for the Lib Dems than the first from the underlying poll data, yet the headline figures are the other way around.

  7. Nick , if you had actually bothered to read my post , you would have seen that the way Comres did their weighting in the 1st of their 2 polls slightly overstated the final headline LibDem figure . Glenrothes will not show anything for either the LibDems or Conservatives as it is a Labour/SNP fight . I think last weeks contests in LibDem/Con fights in Winchester , Lewes and Hampstead gave a good indication as to how the LibDems are doing in their target seats .

  8. I’m unsure whether it is worthwhile looking at the shifts between party in that way – it ignores the shifts in and out of “non-voter” and the impact of likelihood to vote weighting, which can easily change a poll from being good for a party to being bad, or vice-versa.

    The difference in weighting of the Liberal Democrats between the two ComRes polls though is quite ludicrous. 10% of the sample in this poll is people who recall voting Lib Dem in 2005, in the Independent on Sunday poll it was 14.5%, so almost 50% more Lib Dems than the other sample.

    I’ve no idea what they doing, perhaps they use different rolling averages to calculate target weights for the Indy on Sunday and the Independent (which would have no justification I can see – it is very confusing if the same pollster effectively uses two different methodologies). I’m going to have to have a closer look at their weightings.

    It doesn’t seem to be that, but there are some weird weightings there. In one of their July polls they weighted the Conservative 2005 vote higher than Labour’s 2005 vote. If the target weights are that erratic, why bother with political weighting at all?

  9. The Weighted Moving Average is 43:29:18 so not much change. As Anthony notes in his previous blog entry, not much can be read into these polls at the moment. The financial situation is ultra-confusing at the moment for everybody.

  10. Weighting can be all important in the headline published figures . As an example from the US Presidential polls , one company ( Battleground ) has McCain consistently 2% in the lead whereas all other pollsters have Obama 5-7 % in front . On examination it is not the data that is discrepant but the lack of weighting by age group and the weighting for race at 77% white v 73% as per the other pollsters . Battleground too would have Obama in the lead with other pollsters weighting .

  11. Can anyone enlighten me as to whether ComRes use their “generally speaking, do you consider yourself… (Lab,Con,LD,etc)” question as a way of saying something about their sample – the figures are remarkably consistent from one poll to the next, with a sizeable portion non-commital, and more or less neck and neck Lab/Con.

    I think YouGov asked the question once in the last six months, but I’m not sure if that was a one-off, or indeed whether it’s regarded as a valid question in an exercise designed to find out how people would vote to-morrow.

  12. When I said “sample” I meant to say “weighting”

  13. John , Yougov weight by Party ID , Comres ask in every poll the question you mention but AFAIK they do not use the answers for weighting purposes .

  14. Thanks – I’ve always wondered why it’s there – it doesn’t even relate to “how do you think you’ll vote in 2010”-type questioning.

  15. If they ever canvassed me, I’d say “don’t know” as, in common I think with most peop[le, I don’t like being pigeon-holed, despite the fact that I’m just left of centre, rather than just right

  16. Anthony , am I correct in thinking that Yougov ratings for Party ID have remained the same since prior to the last GE ? It is interesting to note that Rasmussen recalculate the US Party ID’s quite frequently in their US presidential polls and do not assume they remain constant .

  17. Barnaby Marder rules….

  18. Mark, they are actually different concepts in an subtle but very important way.

    Rasmussen weight by party ID at the time the survey is conducted, which can obviously change over time. The balance of party ID in the country might be different in August than it was in January, so you’d want to weight to different targets.

    YouGov weight by party ID requested at a point in the past and recorded on panellists’ records, this obviously doesn’t change over time. If the data was originally collected in January, you’d weight to the balance of party ID in the country in January, regardless of when the panel are re-interviewed for surveys.

  19. Thanks for that clarification , Anthony , but then how are people treated who joined the panel recently treated . Yes you can record their Party ID at time of joining but those joining in say 2008 may have a different split of ID’s between parties than those who joined in 2005 .

  20. It’s actually pretty stable – more stable than recalled 2005 past vote – but there are presumably still some differences over time. They used to be excluded from political polls so only people with actual 2005 data were used, now I think Peter has some method of updating targets or accounting for it – you’d have to email him to ask.

  21. Anthony,

    You said it was “pretty” stable over time, which I presume means it does change to a degree. Does the yougov system keep track of how individual members allegience has changed over time, or is it that once updated the previous intention is lost.


  22. As a panel member, I would be interested to be “checked” every so often. I can’t remember how I ID’d myself, but say I was a Lab then, it might well be possible – especially if I moved to Scotland, Peter :) that my ID could change.

  23. Peter – no data is ever lost, all the past surveys are achieved away on the servers somewhere. For detailed things about how YouGov adjust for the slow change in party ID though you’d have to ask Peter.

  24. Mark Senior – a good point earlier, but just to say that the American pollster to whom you referred, Battleground Tracking, has FINALLY noticed that Barack Obama is ahead in the Presidential race in its latest poll.

    I wasn’t aware that I “rule” but I can’t deny that at the moment my Party does.

  25. Actually by that I meant your attitude to sticking to poll discussions… (Sorry)

  26. Yes I realised that. :) I just wasn’t sure whether you were agreeing with my strictures on the subject or you were (shudder, ref the Piranha Brothers) employing sarcasm………

  27. Barnaby , yes Battleground have changed their weightingd for age but are still I believe sticking to 77% whites as opposed to other pollsters 73% .

  28. Slightly off topic here – but is there a site anywhere with historical poll data ? We’re arguing about Thatcher v Foot – one side saying she was saved by Galtieri’s Falklands invasion, the other that she was already ahead in the polls by April 1982.

    Anyone got the figures ? Cos in my memory she was well behind Foot as unemployment and inflation rose.

  29. Laban – Gallup’s monthly polls from 1945 to 2000 are here

  30. Unemployment hit 3 million in January 1982, and continued rising somewhat further. (on the way the figures were then calculated).
    The Liberal/SDP Alliance hit 50% in December 1981, but went straight into menopausal problems, and the polls were quite three way before the invasion.
    The economy had actually started growing in the second quarter of 1981 but it didn’t look like it.
    Interest rates fell steadily in 1982 having been very high from June 1979. There were a few tax cuts in 1982 but not on Income Tax.

    Thanks for those figures above.

    The projected national local elections for May 1982 were C 40, Lab 31, Alliance 26.

  31. It is not my intention to be partisan. But those who are perhaps blinkered by their loyalty to Labour must ask themselves why last year in just ONE MONTH they dropped 7 points and never recovered.

    In September 2007 Labour were averaging 42% but the end of October they were averaging 35%. And it should be noted that at that the point the economy was not the big issue it is now.

    What happened during this period? – The Queen’s speech.
    Brown’s government explained their plans for the future. People had a sense of what Labour were like now under Brown. The polls show that an aweful lot of people thought it sucked! Why?

    Labour were given a chance, but they blew it. You’re right there are still easily enough people out there who instinctively want to vote Labour, but not this Labour.

    I say this as a man from the Rhondda and son of a coal miner who saw the needless hardship he endured as Thatcher closed down his pit which was later proven to be economially sill viable. And why do so many agree with me you should ask?

    Perhaps, one reason is that Labour do not respect people’s freedoms. Perhaps they think they have the right to curtail people’s freedom of speech, such as what people say on websites.

    Even if you do not find the contributions of some not to be meaningful they still have the right to express them. They are not doing any harm. Ignore them if you will. But unfortunately you are typifying the kind of behaviour that puts people off from voting for any party when it gets too big for their boots. On this occasion I happen to think its Labour.

  32. Please excuse the double negative in the last paragraph. I intended to write: Even if you do not find the contributions of some to be meaningful …

  33. Of course not. Nobody deserted the Labour Party a year ago because somebody from Richmond thought that a site about polling shouldn’t be replete with partisan comments cheerleading their own party, or hating another party. The reasons were quite different. It started with Gordon Brown’s dithering about calling the election, carried on with a combination of bad economic news and generally rudderless leadership (with maybe a touch of arrogance thrown in), and culminated in the terrible error of abolishing the 10p tax band. If there is indeed the start of a perhaps small Labour recovery – and it isn’t at all clear yet whether that’s the case – it would be because the Government is showing signs of more decisive leadership, and has regained something of a sense of purpose. I don’t think I’m being unreasonable in wishing for analysis of the polls which appear rather than wanting to read why someone hates the Labour, Conservative or some other party. The idea that Labour went down the pan because one of its members has a vision of what a decent blog about psephology should be like is quite ridiculous, however you dress it up.

  34. Philip J W – actually, people don’t have the right to express those views here. They have to right to express them in general, but this is my website, hosted on a server I pay for. Only I have the right to say what I want on here (within the terms of service of my service provider), the rest of you can damn well follow my comments policy, which Barnaby is rightly defending.

    It isn’t place for non-partisan discussion because Barnaby likes non-partisan discussion; it’s a place for non-partisan discussion because I make the rules and I say so.

    That said, a post exploring the reasons why Labour support dropped in Sept 07 isn’t partisan, it’s exactly the sort of post the site is for.

    That said, your reasons don’t stand up. Labour’s support slumped before the Queen’s speech – the timing suggests it was a result of the humiliating non-election and Brown’s handling of it, the media turning against Brown after his honeymoon and – some people would claim – a successful Conservative conference.

  35. Laban – you aren’t at all off topic, given Cameron’s reference to Thatcher as the change candidate in 1979.

    I found the Tory conference fascinating simply because of the venue. It seemed to have been chosen for the celebration type of conference that should (deservedly) result from a massive turnaround in the polls.

    The fact that all the big speeches were re-written (and presumably largely stripped of oratorical flourish) meant that the mutedness of the performances was stark.

    At his best, Cameron is a small venue performer. A huge speech in a huge venue could easily have re-established that 20 point lead. I’m not sticking my neck out, but I’d suggest any rebound will be smaller for the fact that the venue was no longer helping him.

  36. My guess is that “events” supercede Party Conferences in their effect on peoples voting intention/Polling responses.

    My guess is that Brown’s incredible statement that the Polls were not a factor in changeing his mind to call an Election, was a greater factor in his fall from grace,than Cameron’s 2007 speech-good as that speech was.

    My guess is that current financial & economic events will similarly supercede the effect of Cameron’s speech-good that speech was (IMPO)

    But I don’t know which way current financial concerns will sway people.

    In USA it seems to be towards the opposition.
    In UK it seems to be towards the Government.

    But the GE will probably be in 2010-when events, Conference Speeches, and no doubt policies will all have changed.

  37. Don’t forget the Iraq announcement during the Tory’s conference. I know it’s not scientific but an awful lot of my mates have mentioned that as the reason they went off Labour. I’m not commenting on whether I think it was a deliberate idea by Brown or just very bad timing but I certainly think it had a big effect on public perception.

    I notice that Cameron started his speech talking about the armed forces. I could be wrong but I can’t remember Brown saying much. Does that mean the Conservatives think they’re stronger in this area now?

  38. People are far too concerned about what the next couple of polls will be. We have not reached the watershed moment. That is still a few weeks away.

    Two key events are still to come.

    One is when people stop talking about the ‘credit crunch’ (a phrase not too harmful for Labour) and start to talk about the ‘recession’. The word ‘recession’ has far more impact and will cause more harm to Brown.

    The second moment is Glenrothes. If Labour loses, especially badly, then GB may be in real trouble. He’s been clever with the ‘experience’ argument as it hinders his removal.

    Even if he stays a by-election defeat will kill his momentum and return us to 20% Tory leads just as we offically slip in to recession.

    That will be the water-shed moment.

  39. Surely they can’t really be going to put Alistair Darling in the Home Office, as it hinted in The Economist.

    If David Milliband is sacked, then Jacqui Smith could make a fairly smooth transition to Foreign Secretary.

    Not so good, would be Anne Moffat as Chancellor of the Exchequer, but it would make more sense for Yvette Cooper to take that and have Moffat as Chief Secretary.

  40. I’m not sure whether a Glenrothes by-election loss for Labour is the end for Brown. It will at least partly depend on the extent to which Conservatives and Tories see their votes hold up in what is otherwise a two-way fight.

    I’m also not convinced that a full-blown recession is bad for Brown. I am not satisfied that Cameron and Osborne would do any better in the current situation, so why should I give them a chance? They have yet to explain what they would do differently, and the few gimmicks they have presented do not fill me with any confidence.

    JJB is right that the coming reshuffle will be a good indicator of the feeling inside the cabinet, as overextensive and drastic changes would smack of panic.

    I think Brown should shame Miliband’s conspiring by demoting him, and it would show some backbone to stick with Darling for a bit longer.

    The interesting thing will be the timing of the announcement and how this plays into what I’m expecting to be a lower bounce in the polls for Cameron coming out of the Conservative conference.

  41. Thomas I assume you are referring in your first para. to the Conservatives and the LD’s. In practice their Glenrothes figures are likely to have minimal impact on whether a bad result for Labour is the end for Brown given it is generally accepted it is predominantly a 2 party fight and it would be surprising if one/both of the other 2 parties are not squeezed. I think that Labour could lose Glenrothes badly and Brown would still hang on until at least next Summer.

    I agree with what you say about Brown demoting Miliband but doubt if Brown has the guts to do it. I think you are under estimating Cameron and think Brown will suffer in a full blown recession as he has no room to meaningfully use fiscal policy without adverse political consequences from e.g. a Sterling crises, a down – grading of the UK’s credit rating and much more criticism from the IMF etc.

  42. Can Brown afford to demote Milliband ?

    Just like Blair could not sack or demote Brown – isn’t the same true about Brown and Milliband. If Milliband were demoted he would almost certainly resign and then go looking for 70 signatures on a leadership ballot form which he may not have trouble getting.