Are the election polls always right?
Shaun LowthorpeIf the opinion polls are to be believed then voters will wake up on Friday with David Cameron and the Conservatives the largest single parliamentary party and possibly a slender majority.Shaun Lowthorpe
If the opinion polls are to be believed, then voters will wake up on Friday with David Cameron's Conservatives the largest single parliamentary party and possibly with a slender majority. But are the polls always right?
Six months ago the Conservatives were heading for a landslide victory in the general election, according to the opinion polls.
But then the polls started to narrow. And once the campaign proper got underway, the Lib Dem surge, thanks to Nick Clegg's strong televised leaders' debate performances, has also thrown a spanner in the works, making it harder to know how to call the election.
You may also want to watch:
The polls are pointing to a hung parliament, but there are two great unknowns, which make the result still too difficult to call.
One is the large number of undecided voters seemingly out there (though whether that's because they are genuinely undecided or have made up their minds but do not want to say, could be the crucial factor on which this election turns).
- 1 Man and woman found dead in home
- 2 Neighbours' horror after two people found dead in 'peaceful close'
- 3 Man, 41, charged with Pat Holland's murder as human remains found
- 4 When are GCSE and A-level results out and how fair will grades be?
- 5 Man arrested on suspicion of murder after woman dies in village
- 6 Reward of £20,000 offered after theft of performance car worth £150,000
- 7 Villagers in shock after woman dies in suspected murder
- 8 Hardware store owners retiring after more than 60 years
- 9 Norfolk seaside holiday park battles Shell over solar panel plans
- 10 Woman who bit an officer among eight people arrested in town
The second is that opinion polls are based on the idea of a uniform swing across constituencies from one party to another.
And Nick Clegg's ascent, even if it may now be on the wane, is likely to have put paid to that idea as his party still leads Labour, according to the polls.
Yesterday, Ipsos Mori produced a poll showing that the Tories are now starting to poll better in the marginal seats Mr Cameron needs to win if he is to form a government. And on the same day, Gordon Brown is in Great Yarmouth and Waveney - two of those vital constituencies - on hastily arranged visits to shore up the core Labour vote.
Something is going on, possibly the realisation that the Clegg surge, which seemed to benefit Mr Brown in the early stages, is wearing off, and now looks more likely to split his own vote where he needs it most rather than taking it off the Tories where he doesn't.
A closer look at the polls over the last week seems to be pointing to the Conservatives edging ahead - but in uniform swing land, Mr Cameron needs a 10-point lead over Labour to secure an overall majority and he is nowhere near that.
Labour has now launched its final defensive tactic, which is not about winning the election with a series of gains, but simply not losing it and holding on.
Nick Anstead, lecturer in politics at the University of East Anglia, cautioned that the polls were not always right and the twin factors of voter anger over the expenses and highly regarded local candidates could combine to skew the results in a number of seats.
"The most famous time when the polls got things catastrophically wrong was 1992," said Mr Anstead. "The exit polls were wrong, which is why the BBC and ITV announced it was going to be a hung parliament; by 4am we knew it was going to be a Conservative majority.
"Opinion polls have changed their methodology since then and in more recent years they have tended to be more accurate," he explained. "Some use telephone data, some use internet data and one or two companies do the old-fashioned polling where they go and knock on doors.
"What's most striking is how much they agree even though they have these differences."
He added: "The media are still talking in terms of this notion of a universal national swing, but you are talking about mathematical models that were constructed at a time when we had two-party politics. It doesn't really work. We have got a three-party system. People do not vote predictably with their social backgrounds as they did 40 or 50 years ago.
"All these predictions are based on the fact that voters are going to behave in exactly the same way in different places. That's not how three- party politics or individual constituency politics work.
"A good local MP, a well-run campaign, extra money being put into their area - all these factors can influence the outcome and that's really problematic when you start talking about national swing."
So could that be the biggest irony of all in an election campaign dominated by the three party leaders - that local factors and candidates will determine the final outcome in the minds of so many undecided voters, after all?