politicalbetting.com

Were respondees conditioned to put Ming number one?

With, thankfully, just a few hours to go before the Lib Dem ballot closes the acting leader, Ming Campbell, is now now firmly back in the favourite slot.

All the sentiment today has been in his favour and against Chris Huhne who has been odds-on since the only public YouGov members’ poll three weeks ago put him 4% ahead.

Part of the weakening of the Huhne position has been based on the argument that respondees were “conditioned” into putting the ex-MEP as first choice because of the questions that were put before they got to the voting intention section. These, it was argued, put those being surveyed into a pro-Huhne state of mind.

But did the same thing happened in the earlier poll - said to have been funded by a Campbell-backer which was not published. Information that purports to be from this survey has reached me this evening that, if true, shows that a similar approach was taken with this survey. The only difference is that the conditioning seems to be in favour of Ming.

Although I cannot confirm its authenticity I am publishing it because the person who has sent it has revealed her identity to me although she has asked to remain anonymous. She says that she took part in the survey and saved the pages.

Before the 432 people in YouGov1 filled in the voting intention question they are said to have been asked “How important are each of the following qualities for a leader of the Liberal Democrats?”. They were given five options ranging from “very” to “not” important.

1.Good name recognition amongst the general public
2. Seen as honest and consistent
3. Seen as a serious person
4. Seen as a potential Prime Minister
5. Seen as a potential statesman on the international stage
6. Has a large majority in their own seat
7. A track record of success outside Parliament
8. A track record of success inside Parliament
9. Backing and support of parliamentary colleagues
10. Backing and support of the wider Liberal Democrat party

The critical factor here is that even with these suggested pre-voting intention questions the poll’s finding were not favourable enough to Ming to make his team want to publish the results.

What’s important about both these surveys is that they took place just after the vast bulk of the Lib Dem membership had received their ballots. They give a snap-shot of what opinion was like at the critical time when many of them were casting their votes.

So who has won? I stick with my view that this is too close to call but I’m just inclined to Huhne. Thanks to getting on the MEP after a tip-off from a comment by Alex on the site when the price was 200/1 I make a lot of money whatever happens. I’m not going to risk my large certain profit by going all-in on either Campbell or Huhne.

Roll on Thursday and pay-day.

Mike Smithson

Why I’m sticking with YouGov?

With a sharp move back to Ming Campbell on the Lib Dem betting markets there’s a lot of focus on the one members’ poll that has been published during the campaign - the survey by YouGov commissioned by a rich Huhne supporter.

This showed on first preferences Huhne: 38%: Campbell 34%: Hughes 27%. The run-off figures based on second preferences, eliminating Hughes were Huhne 52%: Campbell 48%.

Much has been made of the fact that the survey was privately commissioned and that the voting intention issue was put last after a series of other questions which it is suggested might have conditioned the response.

When I raised the configuration issue Yougov’s Peter Kellner he responded by pointing to this poll in the Tory leadership contest taken just after the first TV debate.

This was configured in exactly the same way as the Lib Dem survey. After answering a series of questions some of which could be said to “condition” views on David Cameron the Tory members split 68:32 for the younger man. As it turned out Cameron got just under 67% of the votes so this poll, less than four weeks before the ballot closure, came out remarkably well.

It’s very difficult to argue with YouGov’s remarkable success in this niche polling area and only a fool would bet against them. Even though the Lib Dem poll was privately funded it’s the firm’s reputation that is on the line.

Kellner’s view on the February 9 survey was that the contest was “too close to call”. In the absence of any other hard evidence I share that assessment. It looks like Huhne but the gap is too close to regard it as a certainty.

If you want to make money betting on politics you have to detach yourself from your own desires and try to take an objective view of what is going on. As a Lib Dem member I voted Huhne-Hughes-Campbell in that order. But I have arranged my betting so that I win the same from both Huhne and Campbell victories.

The 0400 prices were Huhne 0.95/1 and Campbell 1.18/1. This seems about right.

Mike Smithson

Paul Maggs has devised and will manage the following competition. Simply predict:

1. Who the new leader will be?
2. The percentage of the first preference votes your predicted new leader wins
3. The percentage of the overall votes (after re-distribution of second preferences) your predicted new leader wins

You may make your predictions to one decimal place

Scoring

Your score will be the difference between your prediction and the actual result at both stages of the voting (regardless of whether your predictions are higher or lower than the actual results).

For example, your predictions are Hughes to be the new leader, with 42% of the first preference votes and 54% of the overall votes. The results give Hughes 44% of the first preferences and 48% of the overall votes. Your score on the first preferences is 44 – 42 = 2. Your score on the overall votes is 54 – 48 = 6, so your total score is 2 + 6 = 8.

The entry with the lowest combined score will win the competition.

Note that if your predicted new leader is eliminated after the first preferences, he will score zero in the overall votes. In the example above, if you had predicted Hughes to take 54% of the overall votes but he is eliminated, you would score 54 on the overall votes.

Please post entries in the thread below before 2000 GMT on Wednesday March 1st 2006. Any comments on the competition should be made on the previous thread.

Paul has also devised a prize competion. Details are available here.

Why did Ming’s lead on Hughes 2nd preferences drop 19% in three days?

At this stage in the Tory leadership contest there’d been seven Tory member polls and the only outstanding issue was whether or not the 39 year-old’s share would top the 66% mark.

Contrast that with the Lib Dem race where the only members’ poll in a newspaper was on January 7th - a few hours before Charles Kennedy’s announcement. This showed that Hughes was being beaten by Ming by 49-21. From then on Hughes looked out of it - a view that was reinforced by the statement he had to make about his private life.

The only real question since then was how the Hughes second preferences would split.

Two privately-funded polls by YouGov seemed to provide the answer. One where the fieldwork took place from February 2-6th and another during the period February 7-9th. Although there have been a lot of rumours we have never seen full details of the former. The latter, paid for by a Huhne supporter, was made public.

Although the first preference figures from the Feb 2-6 poll have remained private a poster on this site under the name of “Webpoll2″ did make some information available at comment 27 in this thread. At the time there was a lot of scepticism about its authenticity but the YouGov boss, Peter Kellner, confirmed that it was correct. So of the limited information available we know that:-

Thus the Campbell-Huhne 2nd preference gap from Hughes supporters dropped from 22% to 3% in little more than three days. Could this be plausible? Did the huge change reflect the momentum of the Huhne campaign or could there be specialist polling factors?

The first survey was said to have been commissioned by a Ming supporter and we have not seen how it was structured, what other questions were put and the order in which respondents answered them We do know that in second poll the voting intention questions were put last - and when that happens there is always a suggestion that people’s responses might have been conditioned by earlier answers.

All will become clear on Thursday. This is a close call although I’m just inclined to think that Huhne has it. The latest betting is here.

Mike Smithson

But were too many Labour supporters in the sample?

In an interesting ICM poll in today’s News of the World Gordon Brown is rated against David Cameron on a whole series of measures including who would make a “good Prime Minister”.

In spite of what some of the headlines are saying the question of who would make “best Prime Minister” was not put.

This is not a conventional opinion poll; no questions were asked about voting intention; the only options were Brown and Cameron without any reference to the Lib Dems; and the findings were not, as in normal ICM surveys, weighted by past vote recall.

To a question ICM asked about which party those in the survey had considered themselves to be a supporter of in the past the split was LAB 50.1%:CON 31.5%: LD 17.2%. At the last General Election the three parties got:LAB 36.2%:CON 33.2%: LD 22.7%.

Those surveyed were asked “Putting aside your own party preferences I would like to to think about David Cameron and Gordon Brown. Which if the two do you think….”

Telephone polls like this one involve making randomised unsolicited telephone calls with only about one in six resulting in interviews. For whatever reason this approach almost always produces samples that are heavily pro-Labour. To counteract this in its General Election voting intention polls ICM has pioneered weighting findings on the basis of how people said they voted at the last election. This has not happened with this poll.

So on the final two Prime Minister questions Brown polled 43% and 45% against the 50% of those considering themselves to have been Labour in the past the survey. Thus Brown, it could be said, is not even attracting full Labour support.

This is all very interesting but what we really need are proper voting intention questions along the lines of ICM’s “Suppose the Conservative Party were to be led by David Cameron, Labour by Gordon Brown and the Liberal Democrats by Chris Huhne/Memzies Campbell. If there were to be a General Election tomorrow how would you vote, would you vote Conservative, Labour, Liberal Democrat or for another party?”

Finally - well done to ICM for making the full data-set from this poll available immediately.

Mike Smithson

Is Ming Campbell at 6/4 the value bet?

The seemingly never-ending saga that is the Lib Dem leadership contest takes on another bizarre twist this morning with both the Guardian and Independent reporting on separate surveys of attendees at Thursday’s London hustings meeting.

The challenge for anybody trying to predict the outcome is that the two ex-broadsheet papers come out with completely different findings.

The Guardian spoke to 422 people who were there and found Huhne had 152 first preferences (36%). Ming 124 (29%) and Hughes 87 (21%). Of those they spoke to 59 (14%) were still undecided. There was an almost even split of Hughes second preferences which, if this “sample” is representative would give the race to the former MEP.

But the Independent’s survey talking to 100 people at the same meeting overwhelmingly points to a victory for Ming Campbell. The acting leader had 51 giving him their first preference with 31 to Simon Hughes and just 18 to Chris Huhne.

Neither of these could be described as proper opinion polls because nobody knows whether those who attend a meeting in Central London on a cold February evening are an accurate sample

So with just five days to go before the result is declared this has become a very hard contest to call. The only full poll of party members to be published was funded by a Huhne supporter and was carried out by YouGov from February 7-9th. The only problem with this survey is that the vote intention question was put last and it could be argued that this skewed the result.

Another YouGov survey was carried out privately a few days beforehand and we have only had a partial disclosure of what it found. In a bizarre act confirmed by Peter Kellner of YouGov an anonymous poster on this site sent us the second preferences findings which showed that Campbell was well ahead with the second preferences of Hughes supporters.

While all this has been going on there have been the suggestions that the betting prices are somehow “being managed” - that someone with very deep pockets is using her/her money to maintain the Huhne-Campbell prices within a very narrow range.

My current view: Both Campbell and Huhne have an evens chance. Campbell prices beyond 1/1 are good value and Huhne prices tighter than that level are bad value. I am maintaining an even book between the two front runners.

Betting prices here.

Mike Smithson