Our recent results on the AV referendum, how question wording affects the results, and why referendums are best avoided
Last week my colleague Joe Twyman discussed the results of recent YouGov surveys on the Alternative Vote referendum. These ranged from a 12% No majority, when people are given a brief description of AV, to a 12% Yes lead, after people have been asked a variety of questions about the way we elect MPs and also given a chance to ‘vote’ in a mock AV election.
When people are simply asked the referendum question, without any explanation, recent YouGov surveys have shown a narrow No lead; in common with other polling organisations, we find that opposition to AV has generally hardened in recent weeks.
As Joe said, nobody can be sure which question provides the best predictor of the result of the referendum on 5th May. Indeed, the very fact that the results vary so widely suggests the contest is wide open, for they tell us that many voters have shallow views that may change. When most people have strong views on a subject, poll findings are seldom swayed by the way questions are asked. The great majority of respondents know what they think about, say, taxes or immigration. Few vary their response if a question has some different introductory text, or if they are ‘warmed up’ with preliminary questions on such topics.
The poll showing a 12% Yes lead illustrates the potential opportunity for the ‘Yes’ campaign – but also why they seem unable so far to realise that potential. We conducted the survey for the centre-left think tank, Institute for Public Policy Research. It shows that just 18% of voters are strongly tribal, and ‘strongly opposed’ to all parties other than their own. Fully 77% reject such a tribal approach, and think either that more than one party has some good points (60%) or have little sympathy for any party (17%).
This ought to be good news for the AV campaign, for AV is a system that allows people to express their preferences rather than plump for a single party with an all-or-nothing choice.
Yet when we tested two other features of AV that are logical extensions of the public’s non-tribal approach to politics, the results were less clear-cut.
First, we asked which was more democratic – a system in which each MP needs the support of at least half of all voters, or one in which each MP has more local support than any other candidate. The ‘AV option’ was backed by 45% but the ‘FPTP option’ by almost as many – 37%.
Then we asked whether voters wanted a system in which parties were encouraged to have clear and distinct policies and not to compromise – or one where parties are encouraged to reach out as widely as possible and compromise when necessary. This time the no-compromise FPTP option was backed by slightly more (46%) than the pro-compromise AV option (43%).
Other questions also found that many voters are wary about a system in which their second or third choice might help to decide who becomes their MP.
In short, with just over two weeks to go until referendum day, the ‘yes’ campaign has not yet succeeded in converting widespread support for the general notion of post-tribal politics into a sufficient understanding of, and support for, the way in which AV might allow post-tribal politics to flourish.
In contrast, by concentrating on the simplicity and familiarity of FPTP, the ‘No’ campaign is able to tap into a public mood that is close to torpor.
In a way, our 12% ‘yes’ lead at the end of a detailed poll underscores this point. By the time respondents were asked how they would vote in the referendum, they had been invited to consider a wide range of issues about the way our parties operate and our election system works. I suspect that they were also swayed by the opportunity to ‘vote’ in a mock AV election, and discovered that, far from being complex and mysterious, it was, in fact, easy and perhaps even liberating.
If, somehow, the ‘Yes’ campaign could persuade every voter to consider the issues in the same detail as our respondents, and then go on to convince them of the virtues of 2nd and 3rd choice votes and political compromise, then there would be a very good chance of a ‘yes’ majority on 5th May.
Instead, there is little evidence that many people are sufficiently seized of the issue, or appalled by our present voting system, to consider the proposed change in detail. Indeed, there is a fair chance that – as in most referendums in most countries – a great many voters will have in their mind a different question than the one on the ballot paper. For some it will be: ‘Shall I help Nick Clegg, or kick him where it hurts?’ For others it will be, ‘Do I want to help or harm the chances of an outright Tory victory at the next election?’ For yet others it will be, ‘Whose judgement do I trust more – Ed Miliband and Vince Cable, or John Reid and John Prescott?’
That’s the trouble with referendums: you ask one question and end up getting the answer to another. Which is why I personally side with two figures who have not featured prominently in this campaign – Clement Attlee and Margaret Thatcher – in believing that grown-up politics is best served by avoiding referendums altogether. Parliament should have decided this issue, just as it has decided every other reform in the evolution of British democracy.