Warren in my comments has highlighted a poll by Kindle Research in the Brighton Argus covering the three Brighton and Hove constituencies. The voting intention figures it gives are:

Brighton Pavilion: LAB 26%, CON 16%, GRN 12%, LDEM 5%, WNV 11%, DK 19%, Ref 7%
Brighton Kemptown: LAB 24%, CON 23%, GRN 8%, LDEM 7%, WNV 13%, DK 19%, Ref 4%
Hove: LAB 26%, CON 23%, GRN 7%, LDEM 7%, WNV 12%, DK 15%, Ref 5%

Once it’s repercentaged to exclude don’t knows, refusals and won’t votes, that works out as

Pavilion: LAB 41%(+5), CON 25%(+2), GRN 19%(-3), LDEM 8%(-8)
Kemptown: LAB 38%(-1), CON 36%(+2), GRN 13%(+7), LDEM 11%(-6)
Hove: LAB 38%(+1), CON 34%(-2), GRN 10%(+4), LDEM 10%(-8)

The poll was professionally conducted (ICM’s call centre did the fieldwork) and demographically weighted to each constituency. So why the rather counterintuitive results, especially when set against the ICM poll of Brighton Pavilion at the end of last year?

There are several reasons – firstly, once the don’t knows and won’t votes are taken away there were very small samples sizes of only 200 or so in each seat. Secondly it is implied that the poll didn’t prompt by party, when ICM included the Green party in their party prompt. As I wrote about the ICM poll, this probably helped the Green party and I wouldn’t normally do it, but it was probably justified in the particular circumstances in Brighton. There’s a fair argument against dealing with the unusual situation by not prompting for party at all, but this would have underestimated the level of support for the Lib Dems (and quite possibly the Greens).

Thirdly, unlike ICM the poll does not appear from their presentation and method statement to have factored in likelihood to vote, something which normally favours the Conservatives and penalises Labour. Finally, and most importantly, the poll did not have any political weighting – phone polls without political weighting of some sort will tend to grossly overestimate the Labour party.

With contrasting figures, I would put my faith in ICM rather than Kindle.

44 Responses to “Brighton and Hove poll”

  1. Given the number of clear methodological flaws pointed out above I wonder why they bothered at all.

  2. Remind us of the ICM figures.

  3. Anthony, do you know if we’ll be getting the YouGov/Sunday Times poll tonight? Or will YouGov be working exclusively with The Sun between now and polling day?

  4. @Kieran W: To shape opinion, as opposed to finding it out?

  5. Anthony – I’ve just had this email from Martin Boon – boss of ICM.

    “Nothing to do with us. An organisation called Kindle Research used our telephone centre. I’ve seen it, please advise all it’s not us.”

  6. GIN – yes, I do know ;)

    Neil – the ICM poll is linked from the article, the figures were
    GRN 35%
    CON 27%
    LAB 25%
    LDEM 11%

  7. This is an utterly pointless poll.

    In the GE the greens polled 22%. This says they are now on 12.

    Does ANYONE believe that – in Brighton??

  8. TrevorsDen – no it doesn’t, it shows them on 19%. At least rip it to shreds accurately ;)

  9. Neil – the ICM poll is linked from the article, the figures were
    GRN 35%
    CON 27%
    LAB 25%
    LDEM 11%

    The polls may not be that wrong ! The “preferred” major party of most likely Green supporters is likely to be either Labour or LD.

    Given the national situation and the possibility of a narrow Tory victory, it is not inconceivable that some of these voters are gravitating towards Labour.

    It’s the anti-Tory vote collectively thinking that this is not the time to experiment.

    There is the danger though that the Labour / LD / Green vote could split and the Tories win through with around 30%. However, local anecdotal evidence usually finds the party to back.

  10. Anthony,
    I am a little surprised that Telephone polls are felt to favour Labour so heavily.In the 1983 Election I recall comments to the effect that such polls tended to understate Labour on the basis that telephone ownership was quite middle calss and a sign of relative affluence. I know there have been social changes since that time ,but am still surprised the bias in such surveys has swung so much in the other direction!

  11. Re Graham. We’ve moved on since 1983.

    All the pollsters find that the proportion of Labour voters from last time who are available and ready to answer unsolicited phone calls is considerably higher than the numbers who voted for the party. ICM record an ongoing overstatement of past Labour support of 8% more than actually happened.

    This is why you have to adjust.

    Kindle don’t seem to have done that.

    From what I can see the poll was carried out in such a way that would maximise a Labour response.

  12. Whoops, sorry, didn’t notice the link.

    I notice though that the ICM poll had a sample size of just over 500 – so it’s not wholly reliable either. Much as I agree that it would be good for the country if some more parties were represented in parliament, I dearly hope the Greens are not one of them – they are dangerous extremists in my opinion.

  13. Mike Smithson,
    From what I can see the poll was carried out in such a way that would maximise a Labour response.

    That would certainly explain it. Do you have any explanation for the strange oddities in YouGov’s small Scottish poll published yesterday?

  14. I’d have to agree on the lack of Turnout weighting being the beneficiary to Labour in this polling.

    But what this does highlight is that the Conservatives are winning votes from Labour, as that Labour are losing votes that stay-at-home. And that if Labour can mount a good enough Get Out The Vote campaign, and get their supporters back to the polls, it can all change.

  15. “the Conservatives are *not* winning votes from Labour ” I meant to type.

  16. Graham – it greatly advantages Labour these days. I think because the reasons for skew have changed – back in the early 80s it was still just about a sign of affluence, so could bias a sample towards the better off.

    Since then telephone ownership became almost universal (landline ownership has since peaked and is now declining as people go mobile only), so the skew in phone sampling is now one of availability (are people in) or attitude (will you spend 20 minutes answering impertient questions) – which for whatever reason seems to favour Labour. The response rate in phone polling is now 1 in 12 – so for 1000 interviews, ICM need to ring 12000 numbers.

    Incidentally, the bias in the 1980s from phone polling wasn’t a class thing – since that could be weighted for. The problem was it resulted in an under-representation of the very poorest – weighting up the under-represented DEs just gave pollsters too many of the more affluent DEs. This is the reason for the questions you still get on car ownership – it allowed pollsters to make sure they had the right proportion of the poorest and DEs without cars had a similar profile to DEs without phones.

  17. The Kindle technical notes say that they surveyed 1,000 adults permanently resident in the City of Brighton & Hove.

    Only problem is that Brighton Kemptown has some 16,000 electors who live in Lewes District and not the City of Brighton & Hove. These (according to the 2009 County Council elections) are more likely to be Conservative supporters.

    Kindle also misspell the constituency name. It is Brighton Kemptown not Kemp Town.

    Who are Kindle? And who commissioned this unusual survey?

  18. Tories vote in brighton kempton gone from 39 percent in 1997 to 35 in 2001 to 33 percent in 2005. So it is only a top 30 conservative target seat in name only. Tories will probably do better with target seats above 100 in other areas than in this south coastal area. This was against the national trends from 1997 to 2005.

    In Brighton hove tory vote went from 38 percent in 2001 to 36 percent in 2005.

    In Brighton Pavillion it went from 28 percent in 1997 to 24 percent in 2005.

    Lib dem and green party vote combined has surged from 9 percent to 23 and 24 percent in both hove and and kempton from 1997 to 2005. Greens weren’t even on the ballot in 1997 in Kempton.

    This isn’t a classic marginal seat because the clear trend lines in this district against the national trendlines. These seats will probably be one of the few marginal seats in england where there won’t be a decent swing against labour.

    In Hove tory vote went from 38 percent in 2001 to 36 percent in 2005 against the natiional trends.

  19. These polls just don’t seem right, especially with a 5% increase in the Labour vote in Pavilion.

  20. And yet Kemp Town (note correct spelling) used to be conservative in the 70s. Although the move to LDs and Greens is much stronger here than in most places, the slower change in the same direction nationally says much about why we’ve had Labour in power for 13 years.

  21. Someone talking about a ComRes poll on Political Betting, with figures 40/29/21. Anyone know anything about it?

    Just been on there Andy and saw a very interesting piece about Labour Council destruction having an impact in the marginals. But no Com Res.

  23. The Times is reporting this ComRes poll in a new article today but it seems like the same figures as the last one. Not sure what the Times is playing at.

  24. It’s just their one from last weekend apparently. No idea what the Times are up to there.

  25. By omitting the comma after ‘a future….’, I’m concerned that Labour’s key election slogan ‘A future fair for all’ is little more than promise to give us all some kind of great big street party if they get re elected. Maybe we could include a giant jumble sale as well in an attempt to square the national debt?

    More seriously, Labour really do seem to be up at the moment. On ConservativeHome Tim Montgomerie has been saying Brown’s line about ‘cutting the deficit not the frontline’ should have been a Tory line. It’s peculiar how the side 10 points behind just seem to be so much more comfortable in their mood at present.

  26. Even in my sunniest moments I don’t imagine that Labour can keep Hove, and I don’t think Kemptown (which is often spelt Kemp Town locally) is a serious possibility of a Labour hold either. Pavilion is somewhat different but I don’t have must trust in this particular poll.

  27. @Alec

    The bias of the system leads to the mood of one side being happier.

    The current system means tories need a majority or brown stays in power.

    Brown will not give up power even if the tories have fifty seats more but are less than a majority.

    Getting to 323 seats which would be a majority without Sinn Fein is a huge task.

    It’s all or nothing and if the tories come up just short brown then changes the electoral system which will benefit labour even more the next election.

  28. Just to point out that the ICM poll in Deember was commissioned by the Greens, included the Greens in the party prompt, and was conducted a few days after the Greens had sent a glossy mail shot to every household in Brighton Pavilion. Anthony has given his analysis of that poll earlier, so people can make up their own minds whether that poll was representative, a vanity poll / push poll or whatever.

    Yes, the constituency is officially Brighton Kemptown (though referred to by some now as “Brighton Kemptown & Peacehaven to reflect the geography and better distinguish it from Brighton Pavilion). Kemp Town as a place is a relatively small part of the constituency centered around St James’s Street and the seafront.

    There are demographic changes in both Hove and the eastern parts of Kemptown which may make comparisons with the 70s and 80s difficult, and certainly in my area of Brighton the Lib Dem vote collapsed in 2007 despite a solid campaign, squeezed between Labour and the Greens possibly.

  29. Quite a surprise to see all these three seats going Red- surely these three *precisely* within that category of southern marginals that so many posters on here have convinced themselves are going Blue by leads of 14- 18% (the so called ‘higher marginal swing route to Tory victory’ thesis).

    BUT a poll with questionable methods, samples and questions that might be leading or are framed in a poor social scientific manner.

    We should disregard it whoever it says is in the lead. As we should all polls that can be critiqued in this or other similar ways: like AR.

  30. If Kemp Town is the correct way to write it maybe someone should inform the Boundary Commission! It’s only 60 years since they first created the seat.

  31. Jason

    “Getting to 323 seats which would be a majority without Sinn Fein is a huge task.”

    The SNP has an even harder task in deciding whether they can justify either putting into government the abhorrent Conservatives (whom they hope will be as cack-handed as the Thatcher governments and so help bring about independence) or helping their main, and consistently negative, opponents in Scotland.

    I’m sure they are open to offers, but the price will be high.

  32. Channel 4 reporting new poll with 10% lead for Tories. Rumour is it is YouGov for Sun Times.

  33. Yougov -39/33/17 in Sunday Tims

  34. Tactical voting should have been usefully investigated in Brighton if pollsters had asked who second choice candidate would be in AV system. Brighton, with strong touchy feely greeny groups would have been a good site to poll this. I don’t think that greens are necessarily left as a strong anarchist tendency is also anecdotaly experienced by my own experiences within the enviromental circles in which I have carefully trodden!. But a poll wopuld prove it.

  35. The details of the poll as printed at the top of the page says — “Brighton Pavilion: LAB 26%, CON 16%, GRN 12%, LDEM 5%, WNV 11%, DK 19%, Ref 7%”

    Thats (…let me take another look ….) Green = 12%
    Wikipedia says that in 2005 the Greens polled – 22%.

    Admit did not make it clear I was referring to Pavillion – but with those figures that made sense.

  36. The question that I can’t seem to answer is why did the Argus newspaper run such an unsafe poll …and on the key note day of the Green Party spring conference?

    It seems rather irresponsible and partisan newspaper journalism.

    It doesn’t seem to have been published on their website yet. Perhaps they got cold feet?

    I think it was rather naive of the local paper to think they had valid political poll story.

    If they wanted a poll they should commission a serious ICM -style poll and compare like with like.

  37. The overwhelming consensus is that this is a rogue poll, with no real relevance to the actual reality of what is happening in Brighton Pavilion.

    Warren- Yes, the December ICM poll was commissioned by the Greens but, as Andy points out, the ICM poll was weighted, the Kindle/Argus poll is not. (It hasn’t been pointed out that the poll was comissioned by the Argus, something they neglected to say in their article).

    The Kindle Research poll has a much smaller sample as well, in Pavilion there was only 336 interviewees. After you take out the don’t knows or those who refused to give an answer (37% of the over all poll), you’re actually left with a real sample of 200 people.

    Neither do we have a ward by ward breakdown, nor the full details.

    I think it’s a very silly poll, it’s rogue, and the Greens are quite right to say that the poll comes with a “serious health warning.”

    You can read local opinion on the poll at the following:





    I would include a link to the Argus site with their article on the Kindle poll but I can’t seem to find it, presumably because they were ambaressed by the flawed methodology used and have attempted damage limitation on their credibility for comissioning polls such as this.

  38. Luke – there is no reason to think it’s a rogue poll, since that means something very specific, a poll that produces a dodgy result because of normal sample error, not for any other reason.

    Statistically, 5% of polls will be outside the margin of error of about plus/minus 3% – those are “rogue polls”, and it is no reflection upon the methodology or quality of the poll or pollster, they are inevitable and unavoidable and happen to everyone.

    What a rogue poll *isn’t*, is a poll that is out of line because there is something wrong with it, some other form of error or bias, or because the company used outdated or inappropriate methodology. I hate to be all pedantic, but if “rogue poll” is just used as a term of derision for polls people think have rubbish methodology, we’d need another term for when random sample error just produces odd figures! :)

    (While I’m here, Warren is equally wrong to use the term “push poll”. A push poll isn’t a poll at all – it is a method of direct marketing to participants, where the marketeer pretends to respondents that they conducting a poll in order to spread malicious rumours about an opponent. The tell tale signs of push polls is normally their brevity (they aren’t aiming at a sample, so there is no need for demographic questions, and the aim is to spread the dirt to as many people as possible, so they will be just a minute or two rather than 15 minutes) and the lack of published results (since they are intended as under-the-radar bad mouthing, the people responsible don’t want to draw attention to it!). No proper polling company would ever do it, since its relationship to proper polling is roughly akin to the relationship people who pretend to be meter readers to rob old ladies have to British Gas.

    As far as I am aware there are no instances of it ever having happened in the UK, so you sometimes see people use the term to refer to polls that are biased. Don’t! Save the term up to use should the actual wretched practice of “push polls” ever sneak across the ocean.)

  39. Anthony, thanks for the clarification.

  40. Whilst the Greens are no doubt popular in the Brighton area, it’s important to factor in the local MPs reputations… It’s not just the Co-operative Party connexions that make people like David Lepper are not so easy to oust.
    I think the Greens will probably give Labour a real run for their money in the area, but maybe the best they can realistically do is to come second place in all three Brighton constituencies.

  41. Its really between Tories & Greens in Pavillion. Look at the last couple of times people have voted.

  42. Something seems to have gone rather wrong here.

    I don’t wish to knock it too much because almost any constituency based poll is to be encouraged, and is a rare opportunity to give us a slight indication, but this seems to heavily under-estimate the Greens.

    It perhaps indicates the Con-Lab race is much tighter, low swing in this area, even in Hove, which I can believe up to a point, but would put the Tories favourite in Hove, even with some narrowing towards Labour.

    But polls of course are not there to predict.

  43. Anthony – yes, happy to accept I was wrong to use the term “push poll” but was looking for the right way to emphasise that a poll showing the Greens doing well that was commissioned by them, prompted for them and followed hard on the heels of a blanket mailshot can’t be seen to truly reflect opinion?

    Green Greenie (no doubt who you support eh?) – pointing to votes in European elections (a risk free referendum on the govt/issues of the day) has never been a reliable guide to how a much larger turnout will vote in a General Election when they are picking a government. Thre Greens are not going to form the next govt, even Caroline Lucas would accept that.

    Seal Pup – you are right to mention the Co-operative Party which is very strong in Brighton, but David Lepper is retiring and Nancy Platts is the Labour Pavilion candidate. Simon Burgess in Brighton Kemptown & Peacehaven is a Labour & Co-operative candidate though.

  44. Kindle have provided data tables at http://kindler.posterous.com/full-data-tables-from-bandh-political-party-v following my prompting, and have started responding to comments/criticism at http://kindleresearch.com/blog/?p=276#comment-2955

    I would be interested to know others views of the merits of Kindle’s defence so far.