I’ve been on holiday for the last week, but hopefully haven’t missed too much polling in the August after a general election! One thing that did happen was the Electoral Commission recommending (and the government accepting) a change in the wording of the EU referendum, from a YES/NO question to a REMAIN/LEAVE question. This raises the question of whether or not the wording makes a difference.

At the end of May ICM ran a split sample experiment, asking the then Yes/No version of the question and the remain/leave question that the Electoral Commission ended up recommending. On the Yes/No wording the result was YES 47%, NO 33%, DK 20%; on the Remain/Leave wording the result was YES 43%, NO 35%, DK 22%. Results are here.

ComRes ran a similar experiment at the same time, they asked the then Yes/No version of the question, and a more general question on whether people would vote to stay or leave in a referendum (it didn’t use the exact wording the Electoral Commission have now recommended). On the Yes/No wording the result was YES 58%, NO 31%, DK 11%. On the Stay/Leave question the result was STAY 51%, LEAVE 33%, DK 16%. Results are here.

YouGov haven’t done a split sample, but since the general election they have asked the question in two different ways – one asking the old Yes/No referendum question, and one asking if people would like Britain to remain or leave the European Union. Using the Yes/No referendum question they have found an average YES lead of 8 points. Using a question asking if people would vote for Britain to remain or stay, they have found an average REMAIN lead of 6 points (figures are here, here and here)

The scale of the difference varies between 2 and 9 points and only the ICM poll used the actual question wording. However, the general trend is clear, a remain/leave question seems to produce a smaller pro-EU lead than a yes/no question.

However, what difference the wording makes in an opinion poll is not necessarily the same question as what difference the wording makes in a referendum. An opinion poll is getting someone’s instant reaction having bombarded them with a question they may not have had a firm opinion upon until you asked. A referendum takes place after several weeks campaigning on the pros and cons on each side of the argument and what the implications and consequences of voting Yes or No (or Stay or Leave) might be. I suspect in a referendum, as opposed to an opinion poll, there is very little real difference between Yes/No and Remain/Leave.


ComRes’s monthly telephone poll for the Daily Mail is out today and has topline figures of CON 42%, LAB 28%, LDEM 8%, UKIP 9%, GRN 6%. The last time the Conservatives reached the heights of 42% in any poll was way back in 2010. Full tables for the poll are here.

The poll also had a bank of questions on perceptions of the four Labour leadership contenders and whether people would vote for them. In my last post I wrote about how such questions really don’t tell us very much: none of the candidates are well known to the public, and how people would vote with x or y as leader is an extremely hypothetical question – what policies will they follow? How will the media and public react? Will the party be united, split or riven with dissent? I don’t know and neither do the poll respondents.

For what it’s worth though, the main finding here was how little there was to choose between the candidates. ComRes asked if people would vote for Labour with the four candidates, the spread was between Burnham and Corbyn on 22% down to Kendall on 18%. They asked who had what it took to be PM, Burnham, Cooper and Corbyn were essentially in a three way tie on 23%-24%. Given the standard margin of error on a poll this size, it suggests the wider public really aren’t that enthused by any of them.


-->

With the window for taking part in Labour leadership election closing and ballot papers going out there were several polls over the weekend asking about the leadership candidates, though no fresh polling of people voting in the actual contest. ComRes, Opinium and Survation all had polls asking about the general public’s perception of the candidates. While the polls weren’t presented that way, I’ve seen various people writing about them as evidence of which candidate would actually do better as leader. In particular the Survation poll had Jeremy Corbyn ahead among the public after they were shown video clips, so was taken as a sign that he may not be as damaging electorally as the commentariat widely assume.

Questions about how well different leadership candidates would do in a general election are always popular and sought after, but extremely difficult if not impossible to make meaningful. Asking the general public who they think would do better or worse is perfectly reasonable, but is a different question. Who people think would do better is not the same as who would do better, it’s just asking the public to answer the question for you and a poll is not a Magic 8 Ball. Asking the general public who they prefer doesn’t answer the question either, it contains the views of lots of committed Labour and Conservative voters who aren’t going to change their vote anyway, and preferring is not necessarily the same as changing your vote.

If you ask how people would vote with x, y or z as leader, or if people would be more or less likely to vote Labour with each candidate as leader then you are getting a little closer, but the problem is still that people are expected to answer a question about how they would vote with the candidates as leader when the general public know hardly anything about them. A fair old chunk won’t even know the candidates names or what they look like, the majority will have little real idea what policies they will put forward. None of us really know how they will work out as leader, what the public, press and political reaction will be, how they will really operate. How can respondents really judge how they would vote in a hypothetical situation with so little information? They can’t.

Some polls try to get round that by giving respondents a little more information about each candidate: a run down of their main policy positions perhaps, or in the case of the Survation poll a little video clip of each respondent so people could see what they looked and sounded like. This is better, but it’s still a long way from reality. It’s like the famous market research failure of New Coke – in market research tests people liked the taste, but release it out into the real world and people wanted their old Coke back. A video clip or a list of policies won’t factor in the way the media react, the way the new leader is reported, how they actually handle leading, the way their party and their opponents react. There is no really good way of answering the question because you’re asking respondents something they don’t actually know yet.

Is there anything polls can really tell us about how the leadership contenders would do? Well, firstly I think we can be reasonably confident in saying the polls don’t suggest any of the four candidates is any sort of electoral panacea, the most positive net rating in the ComRes poll is Andy Burnham and just 19% think he would improve Labour’s chances, 14% that he’d damage them (Corbyn gets more people saying he’d have a positive effect (21%) but much more saying he’d have a negative effect (31%). None of them have obvious election-winning magic like, say, Blair did in 1994.

They can also tell us some things about people’s first impressions of the candidates, something that shouldn’t be underestimated (people probably made their minds up pretty quickly that Ed Miliband didn’t look Prime Ministerial, for example, or that there was “something of the night” about Michael Howard. Those early impressions are hard to shift.). On that front the Survation poll is pretty positive about Jeremy Corbyn with people saying he came across as more trustworthy and in touch than his rivals (though such polls are always a bit tricky because of the choice of clips – Survation tried to iron out any potential biasing effect by having clips from each candidate being interviewed on the Marr show, so they were all interviews, all the same setting and same interviewer… but even then you ended up with two candidates defending their position on the welfare bill, one talking about the EU referendum and one talking about rail nationalisation. It’s almost impossible to do such things and have a truly level ground).

The argument against Corbyn isn’t about his personal image and manner though, it’s that he’d put the Labour party in a ideological and policy position that wouldn’t win votes, that the Labour party itself would risk ripping itself apart under a leader with little support among the Parliamentary party and a long history of rebellion. On individual policies I’ve seen Corbyn supporters taking succour from polls showing, for example, that a majority of the public support rail nationalisation or much higher taxes on the rich and drawing the conclusion that there is a public appetite for much more left wing policies. Be careful – look at this YouGov poll which shows a majority of people would support renationalisation of the utilities, increasing the minimum wage to £10 and the top rate of tax to 60%… but also a total ban on immigration and benefits for anyone who turns down a job, making life mean life with no parole in prison sentences and stopping all international aid. There are some policies to the left of mainstream public debate that are popular and some to the right that are popular, it no more means that the public are aching for a far-left political party than for a far-right one. Essentially you can pick a list of appealing sounding policies from almost any ideological stance, from far-left to far-right, and find the public agree with them. In reality though policies require trade-offs, they need to be paid for, they are attacked by opponents and the press. They are judged as a package. In terms of how well the Labour party would hang together under Jeremy Corbyn, polling of the public can’t really tell us – a poll of Labour MPs perhaps!

Bottom line? There is no way of doing a simple poll that will give you a ready packaged answer as to how well or badly a potential party leader will do, and the things that Jeremy Corbyn’s detractors worry about are not things that are easily tested in a poll anyway. My own guess is that those who think Jeremy Corbyn would struggle electorally are correct, though it does depend on whether the Conservatives also pull themselves to shreds after the EU referendum. I am a little wary about arguments about parties not winning because they are too left or too right. While putting yourself broadly where most voters are is sensible enough, those voters themselves don’t necessarily see things as ideologically left and right and specific policies aren’t really that important in driving votes. However, broad perceptions of a party, its perceived competence and the public’s views on how suitable its leader is to be Prime Minister are incredibly important. It will be an extremely hard task for Labour to succeed if it is seem as taking up a risky and radical route, if it’s trying to rebuild a lack of public confidence by selling an approach that is radically different from what a normally risk-averse public are used to, if it is seen as being riven by internal dissent and splits, if their leadership patently doesn’t have the support of its own MPs. Maybe he’ll surprise us, but I wouldn’t count on it.

On other matters, the ComRes poll also had voting intention, their first online VI figures since the election (rather to my surprise. Their online polls for the Independent on Sunday dried up during the election campaign itself and I’d wrongly assumed they’d come to halt as part of ComRes moving their phone contract from the Independent to the Daily Mail. I’m pleased to see I was wrong, and the ComRes/Indy on Sunday relationship continues!). Topline figures are CON 40%, LAB 29%, LDEM 8%, UKIP 13%, GRN 4%, and ComRes have adopted the same socio-economic based turnout model for their online polls that they have started using in their telephone polls.


ICM have released their August poll for the Guardian. Topline voting intention figures are CON 40%, LAB 31%, LDEM 7%, UKIP 10%, GRN 4%. Full tables are here.

This is the first ICM poll since the election to feature an updated methodology in light of the polling error. Since 1993 or so ICM have reallocated people who don’t give a voting intention based on how they say they voted at the previous election. Colloquially this is often known as a “shy Tory” adjustment, as when it was initially introduced in the 1992 to 1997 Parliament it tended to help the Tories, though after the Iraq war it actually tended to be a “shy Labour” adjustment, and in the last Parliament it was a “shy Lib Dem” adjustment.

In practice ICM didn’t reallocate all their don’t knows and refusals as many people who refuse to give a current voting intention also refuse to say how they voted at the last election (ICM call these people “total refusals”, as opposed to “partial refusals” who say how they voted last time but not this time). Under the new method ICM are also attempting to estimate the views of these “total refusals” – they are reallocated at the same rate as “partial refusals” but are assumed to split slightly more in favour of the Conservatives, based upon what ICM found in their post-election re-contact survey. The effect on this change on ICM’s headline figures this month is to increase the level of Conservative support by one point and decrease Labour support by one point.

The implication of this adjustment is that at least some of the error at the general election was down to traditional “shy Tories”, that those who refused to answer pollsters’ questions were disproportionately Conservative supporters. However, from being on panels with Martin Boon since the election and hearing him speak at the British Polling Council inquiry meeting I don’t think he’ll have concluded that “shy Tories” was the whole of the problem, and in ICM’s tables they are clear that they “expect to produce further methodological innovations in the future.”


Tomorrow’s Times has a new YouGov survey of the Labour party leadership election electorate (members, union affiliates and £3 supporters). The figures show Jeremy Corbyn’s lead increasing in the last three weeks – back then he had a seventeen point lead on the first round and just scraped over the line after the reallocation of second and third preferences. The new figures have him comfortably ahead – in the first round preferenes are Corbyn 53%, Burnham 21%, Cooper 18%, Kendall 8%. If the final round ends up Corbyn vs Burnham then Corbyn wins by 60% to 40%, if it ends up Corbyn vs Cooper then Corbyn wins by 62% to 38%. The full tabs are here and Peter Kellner’s commentary is here.

As far as the poll is concerned Jeremy Corbyn is currently solidly ahead (though of course, ballot papers haven’t yet gone out and there is a month to go – indeed, as I write it’s not too late to join Labour and have a vote in the election!). Polling party members is hard, there are not publically available targets to weight or sample too, and there has already been a huge influx of new members and new £3 sign-ups about whom we know little. YouGov’s data has the right sort of proportions of new and old members (thought the final proportions are obviously impossible to know yet), but it’s impossible to know if the sample is right in terms of things like social class. However, looking at the tables Jeremy Corbyn is ahead in every age group and amongst members from every region, amongst working class and middle class members, and amongst members, trade union affiliates and £3 sign-ups.

Corbyn’s least strong group is people who were party members back before 2010 among whom either Cooper or Burnham would beat Corbyn on second and third preferences. People who joined the Labour party between 2010 and 2015 are more pro-Corbyn, meaning amongst all pre-2015 party members the race would be very close. People who have signed up since 2015 are extremely pro-Corbyn, pushing him into the lead.