The Chamberlain Files | Homepage
The early referendum polls, and what to make of them

The early referendum polls, and what to make of them

🕔19.Mar 2012

Forget the result! If respondents in BBC Radio WM’s recent poll on the mayoral referendum are to be believed, the really historic headline on Friday May 4 will be the turnout of 74%. True, most of them, 30 seconds previously, hadn’t realised a referendum was even taking place, but, you bet, they’ll be voting – just like they turn out every year for the local elections.

Back in the real world, less than half that percentage did in fact vote in last year’s council elections. And, by my reckoning, the last time there was a 74%

turnout in even a single Birmingham constituency or ward, never mind across the whole city, was in 1992, when the combined charismatic appeal of John Major and Neil Kinnock attracted record numbers of us to the polling stations.

So does that mean that the Birmingham poll isn’t worth the paper it’s reported on? And likewise a much larger NOP poll for the BBC of Yorkshire’s five mayoral referendum cities – Leeds, Sheffield, Bradford, Wakefield and Doncaster – showing 71% intending to vote? Not at all – for reasons explained later. First, though, a word about those cockamamie turnout figures.

Rarely nowadays do the major polling companies put their intention-to-vote questions, at least during campaign periods, in the unvarnished form that these polls did: ‘Will you be voting in the referendum in May?’. They used to, and would produce similarly incredible numbers. So they started getting respondents to rank their likelihood of voting – definitely, very likely, quite likely, and so on – and analysed the voting intentions of only the ‘definitely’s and ‘very likely’s.

Unfortunately, even the ‘definitely’s usually exceeded the actual turnout, so they switched to 10-point scales – 10 for ‘definitely will vote’, 1 for ‘definitely won’t’ – using the expressed preferences of, in some cases, only the 10s. This was good, in that the proportion of 10s was generally pretty close to the eventual turnout; but bad, because the results have to be weighted to allow for Conservatives’ greater readiness than Labour supporters to say they’ll definitely vote.

There’s no doubt, then, that the implausible intention-to-vote figures do inevitably qualify the message of these polls’ central findings – that, when the idea’s explained to them or they’re pressed to think about it, people on balance like the idea of elected mayors. The question is: do they completely invalidate that message? Why should we believe how these people say they’d vote, when we can be certain that many, if not most, of them won’t vote at all?

Let me suggest two reasons for giving at least some credence to these polls – and to the others that will doubtless follow over the coming weeks. First, as shown in the table, there is a striking consistency in their main findings. The total sample sizes were very different – 500 in Birmingham, 2,500 across the five Yorkshire cities – as were probably other features of their methodologies. Yet the percentages of those saying they’d heard about the referendum, that they’re in favour of an elected mayor, and that they intend to vote, are very similar in the two polls.

(All figures are percentages)


Yorkshire cities



Don’t know



Don’t know

Aware the referendum is happening





‘Do you think Birmingham/all cities should have a directly elected mayor or not?’







Will be voting in the referendum in May







There seems no obvious incentive for these hundreds of randomly contacted individuals to provide similarly misleading responses on a subject in which they have no vested, and precious little other, interest – so why not believe them?


In which case, some of the other statistics from the larger Yorkshire poll also merit attention. As on many issues relating to political perceptions and behaviours, there are considerable age group differences, with 18-34 year olds more in favour of elected mayors than over-55s (59% against 48%), but less aware of there being a referendum (33% against 48%), and less likely to vote.

As to how much an elected big city mayor should be paid, 2% think over £100,000, 65% under £50,000, and 21% between the two – which compares with Councillor Mike Whitby’s current Leader allowances of around £72,000, and the £100,000 proposed by Leicester City Council’s independent remuneration panel for Mayor Sir Peter Soulsby, before he himself chose to axe it.

The second reason for accepting the key figures in these polls as a reasonable representation of current public opinion is that, as noted by Paul Dale (March 12), we’ve been here before. Paul was referring to the City Council’s consultative referendum in 2001, in which voters were presented with the three leadership models then on offer, and almost precisely the same proportion as in the recent poll (53%) selected one or other of the two mayoral options – a majority which, of course, was disregarded by the then Labour council, who preferred the leader-cabinet model supported by the minority 46%.

Birmingham’s, however, wasn’t the only such poll carried out at that time. The Department for Transport, Local Government and the Regions commissioned its own massive survey of approximately 1,100 telephone interviews in each of 10 cities, towns and London boroughs, including the current referendum cities of Birmingham and Bristol, plus Liverpool, which has already decided to ‘go mayoral’ in May – .

Despite the fact that, by the time of the survey in summer 2001, political management reform and mayors had been under discussion for over three years, barely a quarter of the total sample (26%) recalled having ‘heard or seen any information on different ways in which decisions could be made’. No change there, then.

Notwithstanding their limited acquaintance with the issues, almost all those questioned were prepared to express preferences, and 75% supported the option that was supposed to sound like the mayor-cabinet model. For reasons that were obscure even at the time, the actual queston, in addition to being convoluted, omitted any actual reference to elected mayors – asking instead whether the ‘leading councillor should be chosen by existing elected councillors’ or ‘at an election in which everyone in the city can vote’.

Whether or not most respondents realised what they were actually being asked, there was an undeniable consistency across most of the survey authorities (paras. 2.9 – 2.11), with the London Borough of Westminster alone in having more than a quarter (32%) favouring a continuation of the system of councillor election.

Support for direct election was highest in Middlesbrough (86%), a view reflected later in the year in the authority’s binding mayoral referendum, which decisively endorsed a mayor-cabinet form of government. The other figures included Bristol 77%, Liverpool 76%, and Birmingham 75.5%.

Throughout most of the ten years since those surveys were conducted, a dozen or so English councils have had elected mayors. Yet most people, if they’ve not been living in one of those areas, are likely to be in much the same position as they would have been in 2001. They’ve had no personal experience of mayors, and haven’t given them much thought. But now, as then, confronted by a pollster and having the basic principles explained to them, most seem on balance to reckon they’re worth a try.

These are mostly, however, anything but passionately held views, and the dearth of hard information is woeful. The official literature from local authorities will help a little, and we may hear something about prospective mayoral powers in this week’s Budget. But for those who care, either way, there’s clearly everything still to play for.



Chamberlain Files Weekly

Don't miss a thing! Sign up for our free weekly summary of the Chamberlain Files from RJF Public Affairs.
* = required field

powered by MailChimp!

Our latest tweets

Published by

Published by


Our community