Do people prefer Gordon Brown to David Cameron? The answer you get depends on who you ask. There’s a Populus poll mentioned in the Times today (conducted for Opinion Leader Forum - who they?) that claims amongst “swing voters” Gordon Brown would be preferred as Prime Minister to David Cameron by 51% to 24%. I would look very carefully at it before drawing too many conclusions - exactly how are those “swing voters” defined? The Times says they are “people who are inclined to vote Labour at present, but say there is a fair chance of them going to another party, and those currently inclined to vote for another party or unsure about which one to vote for, but who say there is a fair chance of switching to Labour.”
In other words, assuming the Times have accurately described the sub-sample, it is skewed towards Labour - Populus’s normal definition of swing voters is those people who say they might yet change their minds. This does not seem to be that - this sub-sample would include, for example those who aren’t sure but might vote Labour, but not those who aren’t sure but might vote Conservative. It would include Lab/LD waverers, but not Con/LD waverers, etc, etc. The way the Times describes it there are more Labour sympathisers in the sub-sample than Conservative ones, so it is hardly earth-shattering to find they prefer Brown to Cameron.
That doesn’t change the fact that these are largely the people who will decide the next election. People switching between the Lib Dems and Conservatives and not voting at all matter too, but when it comes to the effect of Gordon Brown becoming PM it is the people who might move towards or away from Labour that count. Without a better understanding of the parameters of the sub-sample though it’s hard to say whether it is actually good news for Gordon Brown or not. If those 242 people are mostly current Labour supporters who might yet waver, then the finding that 41% of them think that David Cameron would be as good or better than Gordon Brown could in fact be a worrying finding for Brown. If, on the other hand, a large proportion of those people are not currently Labour supporters, but are other parties waverers (especially if they are Conservatives waverers) then it could be very good news for Gordon Brown. From the information published in the Times we simply don’t know.
(Just for the record, since it doesn’t make that much difference to a 27 point lead, the Times also says “it is statistically a small sample; the findings therefore have to allow for a margin of error of at least plus or minus 3 per cent.” 242 people is indeed a very small sample. It equates to a margin of error of at least plus or minus SIX per cent. Still, even an inaccurate caveat is a step in the right direction!)
This month’s MORI poll has also finally surfaced, this time in the FT, the full voting intentions are not there, only Conservative and Labour. With changes from MORI’s last poll in early September, they are CON 35% (nc), LAB 37% (+1). It’s an intriguing poll. The figures are pretty much unchanged from MORI’s last poll…but MORI’s last poll looked wildly out of place and was seen as a rogue poll, surrounded as it was by Populus and YouGov polls showing Conservative leads of 4 and 8 points respectively. Maybe it wasn’t a rogue poll after all and MORI’s different methodology is just producing divergent figures, or maybe the last poll was a rogue, but this one is picking up a genuine increase in Labour support - perhaps from the veil comments. ICM and YouGov’s monthly polls are due this week, so we’ll soon find out.
18 Responses
It may be that Opinion Leader Forum is not an organisation as such; the phrase usually refers to a forum of opinion-leaders (!), which generally means experts in a particular area, for example an area of medicine. They are opinion leaders because their findings and research lead the way when it comes to the formation of “accepted” scientific opinion.
October 23rd, 2006 at 2:32 amWhat such a forum would look like when it comes to discussing the issue of Brown v Cameron, one can only speculate…
Toby - according to the Times it is a new think tank. We shall see.
October 23rd, 2006 at 10:08 amAs I’m never one to take a single poll seriously I think it is too early to accept that Labour has regained a lead over the Conservatives. However the trend over the last couple of months has been from a Tory lead averaging around 7% in August to around 2% now. Previous changes of government were preceded by good leads in the polls - in the mid 1990s Labour consistently led by over 20% (often over 30%) and in the late 1970s the Conservatives led Labour by 10% plus. On recent poll evidence I can’t see the Conservatives winning.
Incidentally where’s the Lib Dem % in this Mori poll?
October 23rd, 2006 at 10:14 amAnthony,
Could they be connected to “Opinion Leader Research”, who have been doing a lot of stuff for some time.
http://www.opinionleader.co.uk/home.php
Peter.
October 23rd, 2006 at 10:48 amAnthony do you know anything about this poll from Sundays Sunday mail in Scotland, and particularly about, Progressive Partnerships.
Good for the SNP but a sample of only 350 must mean that the margin of erroe is around 5%. Very high list score for the LibDems too….
Peter.
POLL RESULTS
Q1 If a Scottish Parliamentary election was held tomorrow how would you place your first vote for your local MSP?
SNP 35%
LABOUR 32%
LIB DEMS 16%
CONSERVATIVE 11%
SSP 4%
OTHERS 2%
Q2 If the election was tomorrow how would you place your 2nd vote?
SNP 28%
LIB DEMS 26%
LABOUR 25%
CONSERVATIVE 11%
SSP 5%
GREENS 4%
OTHERS 1%
Q3 Who would make the most effective First Minister for Scotland?
ALEX SALMOND 33%
JACK MCCONNELL 24%
ANNABEL GOLDIE 5%
NICOL STEPHEN 5%
NONE OF THE ABOVE 33%
Based on 350 telephone interviews, representative of the general population, conducted by Progressive Partnership from Wednesday to Friday last week.
October 23rd, 2006 at 11:01 amPeter, based on no actual evidence I suspect so, just because Deborah Mattinson had an article in the Times this morning quoting the figures.
The Progressive Partnerships poll appears to be what used to be called “Scottish Opinion” before the last election (I have no idea if they have been bought out, or just expanded into a broader company). I’ve never been quite sure about them - they are a phone based company, but their website suggests they generate their samples from the electoral register. Given that electoral registers don’t have phone numbers on them this doesn’t seem to be possible, even if it is just through data matching I should think that excluding ex-directory people would skew the sample.
October 23rd, 2006 at 11:12 amThe actual figures in the Scottish poll are not that meaningful because of the small sample size . Of more validity though is the difference in voting intentions between 1st and 2nd votes as it is presumably the same people in both votes making a clear statement that they will be voting differently in the 2 votes .
October 23rd, 2006 at 12:33 pmMark,
I’d agree with that. There seems to be a clear shift of people giving second votes to the Libdems, which could have a real effect on the outcome.
The SNP gets a lot of it’s seats on the list, and Labour could benefit from this if it hlds FPTP seats but the list vote gets split between the SNP and LibDems. Depending on the regional make up this could actually save Labours bacon…
We could see Glasgow again all Labour FPTP, but a more even split on the list between LibDem and SNP. Elsewhere we may see Labour list seats falling to LibDems, even in places like the Highlands. You’d have to look at the composition of the lists and see if there is any regional trends.
As I have said before Edinburgh and the Lothians will be the most interesting battleground, and we could see for the first time, Labour as a regional force, with limited control or support beyond West Central Scotland.
Peter.
October 23rd, 2006 at 1:34 pm2003 Breakdown for Scotland was,
Labour ;- FPTP ;- 46 List;- 4
SNP ;- FPTP ;- 9 List;- 18
Lib Dem ;- FPTP ;- 13 List;- 4
Tory ;- FPTP ;- 3 List:- 15
SSP ;- FPTP ;- 0 List;- 6
Green ;- FPTP ;- 0 List;- 7
Shares of the vote in 2003 were,
Labour ;- FPTP ;- 35% List;- 29%
SNP ;- FPTP ;- 24% List;- 21%
Lib Dem ;- FPTP ;- 16% List;- 12%
Tory ;- FPTP ;- 17% List:- 16%
SSP ;- FPTP ;- 6% List;- 7%
Green ;- FPTP ;- 0% List;- 7%
If there really is a list surge for the Libdems, then it may be the Tories that get hurt most.
Also the poll hasn’t yet seperated the SSP from Solidarity, and it looks like neither will have FPTP candidates. On these figures the Greens and SSP may well be loosing out to others, in part because of the whole sheridan fiasco, and the fact that almost everyone in Scotland is going green.
Peter.
October 23rd, 2006 at 1:55 pmThe LibDem figure in the Mori poll was 18% and Others 10%
October 23rd, 2006 at 2:56 pmAll very odd - seems from the Mori website that the non-certain to vote are split 46% Lab 28% C. Can this be true? I guess we’ll see from other polls if this is a rogue.
October 23rd, 2006 at 5:13 pmLooking at the WMAs the last Mori poll understated the C lead by 6.1 points (and 7.6 on an interpolated basis) and Mori polls have the highest Standard Deviation from the WMA and the Interpolated of any organisation (3.2/3.1 resp.)
October 23rd, 2006 at 5:18 pmIt’s interesting how MORI seems to consistently give Labour a better rating compared to other polling companies, and the Conservatives a less favourable one.
Some people have been saying they must be getting it wrong somehow, but we shouldn’t forget that MORI got the exit poll of the last election absolutely spot on.
October 23rd, 2006 at 5:21 pmIf I remember correctly, MORI got the exit poll spot on because they made changes to their methodology for the eve-of-poll polls to bring their figures in line with the other pollsters.
I know this isn’t the right thread, but I just wanted to say a quick thanks to Anthony for the Iraq Deaths post - very interesting reading and as usual a completely impartial and thorough analysis. And you seem to have managed to tackle the subject without attracting too many loony commentators!
October 23rd, 2006 at 5:32 pmThere are separate things here - MORI’s last eve of poll prediction was well within the margin of error of the actual result, but used a slightly different methodology from their normal methodology (this isn’t unusual, MORI have always made extra adjustment to their eve of poll survey on the grounds that it is an actual prediction, unlike any other polls.)
The exit poll was a joint venture between MORI and NOP. Exit polls are entirely different creatures from other polls, the sampling method is entirely different, one needs to take no account at all of likelihood of voting (if they are coming out of a polling station they voted!).
October 23rd, 2006 at 5:39 pmThe final MORI election poll in 2005 had only two very minor adjustments that were not applied to all our other election polls; these accounted for the likely votes of those who refused to say how they would vote and for an anticipated higher “overclaim” of voting certainty among Labour and Lib Dem than Tory supporters. Each of these adjustments changed one party’s percentage support by one point and left the others unchanged.
October 23rd, 2006 at 5:44 pmIn fact MORI’s final poll was less adjusted and weighted than any of the other companies’ final polls. The raw unweighted and unadjusted figures (excluding those expressing no voting intention at all) were Conservative 31%, Labour 38%, LD 23%; the final published projection was Conservative 33%, Labour 38% and LD 23%. No fiddling, no manipulation, just good sampling.
There’s an explanation of the OLF sampling here:
http://theprogressive.typepad.com/the_progressive/2006/10/brown_leads_cam.html
I’d be interested in your views. Does this make for a fair cross-section of “swing voters” or is it a just ropey sampling?
October 24th, 2006 at 6:14 pmPC - I don’t think it does no, or at least, it could make a for a decent cross-section but we don’t have the necessary information to judge. Swing voters overall include those moving in and out of the Labour camp, and those moving in and out of the Conservative camp, the Lib Dem camp and so on. Some of the these groups intersect with one another, but overall this sample is only those moving in and out of the Labour camp, so it should hardly come as a great surprise that they favour the likely Labour leader.
In a question comparing Brown and Cameron like this, I would have included either all swing voters, or all those voters who say they might potentially back both Labour and the Conservatives. By including, for example, those who are swinging between Labour and the Lib Dems, but not those swinging between the Conservatives and the Lib Dems, you’ve got a sample that is far more favourable towards Brown than Cameron.
October 25th, 2006 at 9:21 am