Professor Colin Mellors: Polls tell the story of struggle for power

CAPTURING the swing voters – the "undecided" and those living in marginal constituencies – is key to winning any election. Although 45 million people will be eligible to vote on May 6, a much smaller number will, in reality, determine who enters Downing Street.

And, as David Cameron demonstrated on Tuesday by visiting Leeds for his first day of campaigning, West Yorkshire can expect to see a steady stream of political big hitters from all of the parties in the coming month.

The electoral arithmetic indicates the scale of the task facing Mr Cameron. Even though it is estimated that his party benefits by a dozen seats because of the new boundaries, he still needs a further 116 seats for outright victory.

Hide Ad
Hide Ad

A swing against Labour of 1.5 per cent is needed before they lose their majority, one of 4.3 per cent before Conservatives overtake Labour as the largest party, and a massive 6.9 per cent for an overall majority – larger than that achieved by either Ted Heath or Margaret Thatcher when they came to power.

Crucial to the campaigns will be the work of the opinion pollsters, for whom elections are big business.

Each day, their findings – especially the private polls commissioned directly by the parties – will be scrutinised in minute detail for hints of what is going well and what is going badly.

Polling, of course, is not an exact science, as was demonstrated so spectacularly in 1992 when the main polling companies credited Labour with a 0.8 per cent lead which would have put Neil Kinnock in Number 10. In the event, John Major won with a 7.5 per cent margin over Labour; the polls had been a full eight points adrift in

their predictions.

Hide Ad
Hide Ad

The failure was ascribed to a mixture of sampling error, an unexpected late swing, and the high proportion of "shy" Conservative voters who refused to volunteer their intentions.

Even though techniques have improved, the headline poll figures that will feature in the media for the next four weeks – there were 46 during last month alone – have still to be treated with a degree of caution. Different interviewing methods (telephone, internet, panel, face-to-face), sample sizes, sampling methods, coverage (nationwide or marginals), and weighting can all affect the calculations.

It helps explain why, five days ago, one poll showed a 10-point

Conservative lead while another reported it down to four points. The former put the Conservatives within sight of victory while the latter would leave Labour as the largest party in a hung Parliament.

Hide Ad
Hide Ad

Polls are simply snapshots of opinion, and using them to predict the outcomes for individual constituencies is hazardous. It assumes a uniform swing (which is rarely the case) and largely ignores the implications of three-way contests (and the impact of local issues).

There is the added complication of new boundaries – only four remain intact in Yorkshire and Humber – which means that pundits will be

basing their calculations on "notional" results from 2005 and the chart that will appear in the Yorkshire Post each week uses these notional results as its basis.

Throughout Gordon Brown's premiership, apart from a brief period in autumn 2007, the Conservatives have been comfortably ahead. Last April, for example, the Conservative lead averaged 10 per cent, equivalent to a swing of 6.5 per cent and close to what is needed for an overall majority.

Hide Ad
Hide Ad

Local election results have provided the Conservatives with even more cause for optimism, with leads over Labour of 14 per cent (2007), 18 per cent (2008) and 13 per cent (2009) respectively. On these figures, all but the safest of Labour seats would have been vulnerable.

However, since January, things have changed. While last October polls indicated a Conservative lead at, or above, 14 per cent (8.5 per cent swing), more recent ones suggest a hung Parliament, albeit one with the Conservatives as the largest party.

So, with four weeks to polling day, where do the parties currently stand?

The last 10 days have helped the Conservatives, with the spat over National Insurance contributions prevailing over Labour's attempts to highlight contradictions in Conservative spending plans. Their lead is close to double figures, although Liberal Democrat support has also hardened in some polls (again, seemingly, the result of the "Vince Cable" effect). Although doing well, the Conservatives remain tantalisingly short of the 10 per cent margin that is necessary to win outright.

Hide Ad
Hide Ad

Projected uniformly, Conservatives would be 39 seats ahead of Labour but still 25 short of an overall majority. The Labour seats at risk in this region are shown on the chart and, on these figures, Conservatives could also expect to pick up the new York Outer seat.

One of the more intriguing findings is that, according to one poll,

over one third of voters might shift their allegiance if they believed a candidate was tainted by last year's expenses affair. That, combined with the large number of undecided voters, suggests that many voters have still to be won over.

It is conventionally believed that governments lose elections rather more than oppositions win them. Given what seemed an unassailable position just a few months ago, the Conservatives will be hoping to ensure this will not be one occasion when the opposition loses an election. Each week, until the election, the Yorkshire Post will be tracking the opinion polls to gauge the mood of the electors.

Professor Colin Mellors is a political scientist at the University of York.