ICM show YES campaign ahead in AV vote

ICM have a new poll out for the Electoral Reform Society about the AV referendum (full tables here). Asking about voting intention in the referendum 35% say they will vote yes, 22% no, 9% would not vote and 35% don't know.

Regular readers will no doubt notice the stark difference between this and the regular YouGov tracker poll on AV voting intention, which on it's last outing a week and a half or so ago had the NO vote ahead by 41% to 35%, with 7% not voting and 17% don't knows.

The big differences are in the level of don't knows - ICM has twice as many as YouGov - and in how Labour supporters say they will vote. In ICM's poll Labour identifiers (they didn't ask voting intention) are in favour of AV. In YouGov's polling Labour supporters, initially supportive of AV back in May and June, have gradually changed their opinion and are now against it.

The contrast definitely won't be down to the differences between telephone and online polling, which is one of the things I've seen floated, since the ICM poll was also conducted online, presumably using their own panel. The difference is more likely in the wording of the two questions - there are three possible differences here:

1) The big difference is that YouGov have introductory text explaining briefly what the two systems are (the actual explanations YouGov use originate from a a survey for the Electoral Reform Society back in November 2009). ICM's poll doesn't tell people anything about the systems.

2) ICM actually don't mention FPTP at all in their question, it asks if people would vote for a new system called the alternative vote, or to keep the existing system. YouGov present it as a choice between two systems.

3) In their preable YouGov mention the referendum as something the coalition government are doing, while ICM just say it is due to happen in 2011.

Now, we know from various other bits of polling that many people have very little knowledge of what AV is - the Electoral Commission's qualitative research commissioned as part of their evaluation of the referendum question found people thinking the alternative vote meant being able to vote online or by post, referred to an unspecified alternative, or was a proportional system. I think it's reasonable to assume that a lot of the reason for a lower level of don't knows in the YouGov poll is that respondents have been told what they are voting on.

I'm more intrigued about the second difference - how come ICM's Labour identifiers back AV but YouGov's Labour voters oppose it? Perhaps it's because YouGov mention that the referendum is a coalition policy that makes Labour voters less supportive.

All that aside, the question people will probably be interested in is which one will give the right steer to the result. Normally my instincts are to give respondents the absolute minimum level of information they need to respond to a question, so an ICM style question has its attractions (I wouldn't actually use the same question ICM did, but that's neither here nor there). The big difference wording would appear to make suggests public opinion is still unformed on the issue, and ICM's finding that 35% of people don't know how they'll vote if you ask them straight out tells us something important about how many people don't know about this.

Equally the fact that when YouGov ask people's intentions having told them what AV is they are far more negative doesn't bode well for the YES campaign. As the referendum gets closer people will become more aware of the choice and telling people about FPTP and AV should make less and less difference to their answers. Nevertheless it would be naive to assume that by the time of the election everyone voting will know what AV and FPTP is - the electoral Commission will send all voters a booklet before the referendum telling people what AV and FPTP are - but that doesn't guarantee everyone will actually read it.

I'd also expect it to be seen as more of a choice between two system, and the political context of the coalition being an important factor - so my guess is that the YouGov question gives a better steer to how things will actually turn out... but my experience is that most people will want to believe the poll that tells them the result they'd like, so feel free to do so. The lessons I'd take away from the polling is that lots of people don't know, increased awareness appears to help the NO campaign (something YouGov polls prompting people with pro- and anti- arguments and reasking VI have also shown), and that the YouGov trackers suggest that the trend so far has been towards no (probably due to increased animosity towards the coalition government).

Going back to the ICM poll there are also questions on whether people would be more or less likely to vote Labour or Conservative if they backed AV. Regular readers will know my general disregard for questions like this (I have the full old rant here). They give issues of little salience false prominence, respondents tend to use them to indicate approval or disapproval for a policy regardless of whether it will actually change their vote, and people saying it would make them more likely to vote for X are invariably mostly people who are already voting for X.

In this case, 13% said backing AV would make them more likely to vote Labour, 10% less likely. ICM didn't ask voting intention - but they did ask party ID. The majority of people who said they'd be more likely to vote Labour if they backed AV are Labour identifiers anyway, the majority who said it would make them less likely to vote Labour are Conservative identifiers anyway. Turning to the Conservatives, 8% said they'd be more likely to vote Conservative if they backed AV, 13% said it would make them less likely. Just as meaningless as the Labour question, but kudos to the ERS for releasing questions that probably didn't get the answers they'd have liked. I admire pressure groups that do that.