A new ICM poll in the Daily Mail shows 82% of people think the revised EU treaty should be ratified by referendum, rather than left to Parliament. As David Boothroyd rightly pointed out in the comments here over the weekend, that doesn’t necessarily mean much as polls invariably tell pollsters they like referendums on nearly anything you care to name: it is almost the equivalent of asking “do you think your opinion on X should count?” – how many people’s answer will be “Crikey no, I’m a complete numbskull, don’t give me a vote”.
Potentially more interesting are questions measuring whether people actually care enough about the issue to change their vote. The ICM poll found that 24% of present Labour voters said they would be less likely to vote Labour at the next election if they decided not to have a referendum. 13% of Labour voters said it would increase their likelihood of voting Tory at the next election if they promised a referendum.
If the issue was really likely to lose Labour a quarter of their support this would be massive news…except it isn’t. It’s just an idle threat. Saying you’ll be less likely isn’t the same as saying you absolutely definitely wouldn’t, and telling a pollster you might not vote for a party if they do (or don’t) do something you want is an easy, cost free, hit. It’s not the same as actually doing it.
A EU referendum could potentially be a significant issue in an election, but questions like this can’t really tell us how much.
Since this poll gave splits for Labour and Conservative voters, it implies that a proper voting intention question was asked. We’ll see what else pops up when ICM put the full results on their site.
UPDATE: The full results are now available on ICM’s website here. It doesn’t contain any voting intention figures. The website says that “The vote intention cross break in this poll was included for analysis purposes only. ICM’s usual vote intention questions were not asked.” There are cross break figures for voting intention, but the poll wouldn’t have contained things like the likelihood to vote question and it’s possible it didn’t even use the same wording, so you can’t use them to engineer figures comparable to a proper ICM voting intention poll. Hopefully ICM’s monthly poll for the Guardian should be along soon anyway.