If there is a second referendum, there is no good reason to suppose that it will deliver the result that remainers hope for. I’m basing that view not on opinion polls, but on some old-fashioned political science. There is no such thing as ‘the will of the people’. What there is, instead, is a mish-mash of different opinions. Some people voted ‘leave’ because they were unhappy with the EU; some because they were opposed to immigration; some because they were against capitalism; some because they wanted to return to the 1950s; some because they wanted to give the government a kicking. Some people voted remain because they like the EU; some because of self-interest; some to avoid disruption; some because of their judgment about the economy; and so on. Lies or fear may have played a part, on either side, but that’s not decisive; nor is the fact that some people will feel empowered to vote leave, or that other people will strain themselves to get a different result this time. The more complex an issue is, the more likely it becomes that people with different motivations and preferences will cancel each other out, and the closer the result moves to what you’d expect from a random distribution – a 50-50 split.
Once we start from that position, the result is statistically likely to be decided by a relatively small group of people with a strong, settled opinion, if there is no equivalent group on the other side to oppose them. The source of this argument is L Penrose, The elementary statistics of majority voting, Journal of the Royal Statistical Society 1946. Bartholomew and Bassett wrote, in Let’s look at the figures, that “2,000 resolute voters in a population of just over one million can almost always get their way.” (p 125) And that’s what happened in 2016. (There might well have been an equivalent group on the other side – Britons in Europe – but they were largely barred from taking part.) It’s not the polls that count; it’s the mechanism by which the issue is to be decided. And without very strong reasons to the contrary, we should expect the same mechanisms and the same process to produce the same result.