The front page of the Evening Standard today has as a big headline of “Boris to be Mayor Again”. It is a rather brave conclusion based on this ComRes poll.

Even being generous, it isn’t really a voting intention poll – it’s asking who people’s favoured candidate would be out of Boris, Ken and Oona King, with Boris in a healthy first place.

Naturally this did not include a Liberal Democrat candidate, not candidates from other parties, but that’s acceptable given that the London Mayoral election is conducted under a preferential system and the final run off will almost certainly be between Boris and the Labour candidate. Including Ken and Oona King is rather odd though, given it splits Labour support between two candidates – it might have been better to do two run off questions. Still, in this case it thankfully does not give a misleading result, since even if you add Ken and Oona King’s preferences together Boris would still be more popular than them.

There are further question marks here though. Over on Liberal Conspiracy Don Paskini raises some sensible points about question order. Voting intention questions are nearly always asked first in an opinion poll, and certainly any questions that risk skewing answers are avoided. In this poll people appear to have been asked whether they agreed with Boris that it should be made harder to strike on the underground, and that London should be protected from cuts… then asked who was their favourite candidate (I should add that Don’s allegation that the poll was deliberately skewed to help Boris is almost certainly nonsense. No professional pollster would do such a thing – we have some integrity you know.)

Meanwhile Mike Smithson on Political Betting points out that the poll does not appear to have been politically weighted. ComRes always weight their voting intention polls, but sometimes don’t do it on other polls, where they claim it doesn’t make much difference. Even if Mike’s correct, it’s unclear what difference this would have made – political weighting of phone polls helps the Conservatives, but this was conducted online. In ComRes’s previous published online poll political weighting helped Labour.

So, what is the actual position? Well, the last actual mayoral voting intention question I can find is over a year old from YouGov in April 2009, which found in a re-run Boris would beat Ken by 49% to 33%. A lot of water has passed under the bridge since then, but despite all the question marks mentioned above it would be churlish to say this poll isn’t pretty positive for Boris. However, I wouldn’t be confident enough to treat it as a voting intention poll. Nor would I conclude that Boris is set to be mayor again when there is a year and a half of campaigning and massive government cuts to come before any actual votes are cast.

UPDATE: YouGov’s daily GB voting intention figures for tonight are CON 42%, LAB 38%, LDEM 11%

In advance of the Liberal Democrat’s first party conference in power YouGov surveyed 566 Lib Dem party members. On the whole members remain strongly supportive of the Government, Nick Clegg and the decision to go into coalition, but express concerns over economic policy, reject the idea of a pact with the Tories and, in an ideal world, would prefer a deal with Labour. Here’s a summary of the main findings:

58% of Lib Dem members approve of the Government’s record, with 23% disapproving. David Cameron’s approval rating is +64, Nick Clegg’s +70 and Simon Hughes’ +51.

Only 29% of party members fully agree with the Government’s policy of cutting spending to reduce Government borrowing. 28% would like to see borrowing reduced more gradually, and 35% think the Government are right to reduce borrowing quickly, but would rather there were higher taxes and fewer cuts.

78% of party members approve of the decision to enter into coalition with the Conservatives. Asked what the party should have done given the circumstances after the 2010 election 50% think a Conservative coalition was the best solution, 22% would have preferred a deal short of a coalition, 19% would have preferred a deal with Labour.

But in a hypothetical hung Parliament situation where the Liberal Democrats could form a majority government with either party, and both offered equally good deals, 46% of Lib Dem members would prefer a deal with Labour. Only 26% would go with the Conservatives.

A narrow majority of Lib Dem members (53%) expect the coalition to last the entire length of the Parliament.

Lib Dem members reject the idea of an election pact with the Conservatives at the next election by 21% to 66%.

On a left right scale 65% of Liberal Democrat members identify themselves as being left-of-centre, with an average score on a scale of -100 (very left wing) to +100 (very right wing) of -32. Nick Clegg is seen as more centrist, with a score of -7. Deputy leader Simon Hughes is closer to the average of party members, with an average score of -42.

This article is cross-posted from the Yougov website here, the full tables are up here.


As far as I know there are at least two new polls tonight. First up, YouGov’s weekly poll for the Sunday Times has topline figures of CON 41%, LAB 39%, LDEM 13%. There will likely be other questions once the Sunday Times itself goes up online, but I will update those later.

There is also a ComRes poll for the Independent on Sunday. Topline figures are CON 37%, LAB 35%, LDEM 15%. There are no changes, since the poll was conducted online, so by definition will have had completely different sampling from ComRes’s usual telephone polls (as far as I can tell from the tables, the rest of ComRes’s methodology is identical to that used in their telephone polls.) ComRes’s website suggests they will be carrying out concurrent online and telephone polls, rather than switching from telephone to online.

Ipsos MORI’s political monitor for September is out and has topline figures of CON 37%(-3), LAB 37%(-1), LDEM 15%(+1). Changes are from MORI’s July political monitor – there does not appear to have been one in August.

This is the second poll to show the parties equal after ICM’s last month. YouGov earlier this week showed the two main parties within one point of each other. YouGov’s voting intention this morning meanwhile was CON 42%, LAB 39%, LDEM 12%.

Something interesting for the methodology geeks amongst us – below is a graph of past vote recall for the 2010 election from MORI’s five polls since the general election. Unlike most other pollsters do not weight by past vote, so while we should expect their past vote recall to be more volatile than other companies, any trends that are there will not be weighted out of existance – note the percentage of those saying they voted who say they voted Liberal Democrat. In May, 28% of people telling MORI they voted at the election said they voted Lib Dem, clearly higher than the Liberal Democrats actually got. In the months since then, the proportion of people claiming they voted Lib Dem at the 2010 election has dropped to 21%. We’ve gone from people claiming they voted Lib Dem when they probably didn’t, to people conveniently “forgetting” that they ever voted Lib Dem.

YouGov’s daily poll for the Sun tonight has topline figures of CON 40%, LAB 39%, LDEM 12%. It’s the lowest lead YouGov have shown since the election (and, indeed, since the election-that-never-was), although ICM have already shown the two parties neck and neck.

We have been seeing the gap between Labour and the Conservatives gradually narrowing over the summer. We are rapidly heading towards the conference system, which looking at past elections has normally produced a lot of up and down in the polls as each party gets some sort of boost from the media coverage of their own conference – I’d be surprised if we didn’t get a poll showing Labour ahead during their conference if we don’t get one this week.