Do voters want final say? What to make of ICM poll.

Are British voters now clamouring for the right to decide whether Brexit goes ahead after all? A new poll, by ICM for the Guardian, suggests a groundswell of support for a new referendum in due course. For those who favour this, the news is certainly encouraging. But let us not – yet – go overboard.

ICM asked people whether they agreed or disagreed with this statement: “I think the public should have the chance to take a final decision on whether or not to leave the EU in another referendum when the outcome of the negotiations is known”. The responses: 47% agree, 34% disagree. The remaining 19% say “neither” or “don’t knows”. Counting only those who take sides, the figures are: agree 58%, disagree 42%. This gives today’s Guardian its front-page headline: “Surge in support for a second EU referendum”.

This poll deserves to be taken seriously for two reasons. ICM is one of Britain’s most reputable polling companies; and its sample size, more than 5,000, was exceptionally large. Their figures cannot be explained by some sampling freak.

There are, however, two reasons for caution. The first is that the Guardian’s phrase, “surge in support”, implies that ICM have found a movement in public opinion since it last asked this question. However, I can find no past ICM poll in which this question appears, and the Guardian does not cite one to support its headline.

Secondly, there is mounting evidence that significantly different results can be obtained by asking the new-referendum question different ways. Michael Ashcroft recently commissioned a survey that tested this point. He does not name the company that did the fieldwork, but the layout of the tables on his website indicates that the work was done by YouGov, another highly reputable company (full disclosure: I used to be its president). Ashcroft’s fieldwork (January 11-14) was much the same as ICM’s (January 10-19), so we can compare the two surveys directly.

Ashcroft divided his 3,300-strong sample into four different groups of 800+. Each was asked a different version of the new-referendum question. When people were asked simply: “Do you want a second referendum on Brexit?”, they responded 38% yes, 51% no, 11% don’t know. Excluding don’t knows, these figures are 43% yes, 57% no.

When a different, statistically similar, group was asked: “Once Brexit negotiations are complete, would you support or oppose holding a referendum on whether to accept the terms, or leave without a deal?”, 39% supported the idea, while 31% opposed it and 30% were don’t know. Excluding don’t knows: 56% support, 44% oppose.

So, differently worded questions produced figures ranging from a 14% majority against a new referendum (when don’t knows are excluded), to a 12% majority in favour. (Ashcroft’s two other versions of the question produced results between these two.)

One can argue which is the “correct” way to word this question. The real point is that when different phraseologies produce such different responses, it is a sure sign that many voters are not thinking much about the matter; this means that when asked a question out of the blue, their reactions differ according to the precise way in which the question is put.

My conclusion from the ICM and Ashcroft polling evidence is that the campaign CAN be won for a people’s vote on the final Brexit deal (or non-deal). If the campaign  frames its case effectively, many voters will respond positively. But the views of too many voters are too shallow to claim victory just yet.