|
Poll analysis: Friday April 22
MORI chairman Sir Robert Worcester analyses the latest opinion poll data.
I must have heard or read the words 'in 2001 the polls overestimated the Labour vote' scores, if not hundreds, of times over the last four years.
These comments have reached a crescendo during the first fortnight of this election. Few of these commentators and interviewers have acknowledged that we have taken steps to take account of the lower turnout among Labour supporters in recent elections. But to what effect?
The Sun newspaper reports the latest MORI poll, taken among a national sample of 1,001 electors by telephone. The voting intention figures reported on those who said they were 'absolutely certain to vote', 593 electors or 59 per cent of the total (which happens to be the turnout in the 2001 general election). The results were Labour 39 per cent, Conservatives 32 per cent and Liberal Democrats 22 per cent.
Among those giving a voting intention, whether 'certain' of not, (722 people) the shares were Labour 41 per cent (+2), Conservatives 28 per cent (-4) and Liberal Democrats 22 per cent (no change).
The cause of the difference? Three in four (75 per cent) Conservative supporters said they were 'absolutely certain to vote', while fewer (62 per cent) Labour supporters were so sure, as were 67 per cent of Liberal Democrats.
In 2001, MORI reported 45 per cent Labour, 30 per cent Conservative in our eve of poll survey. If these correction factors had been applied in that election, Labour would have been reported to have 43 per cent, the Conservatives 34 per cent and the Liberal Democrats 18 per cent.
The result on the day was in fact Labour 42 per cent, Conservatives 33 per cent and the Liberal Democrats 19 per cent.
One point out for each party, well within the usual +/-3 per cent quoted for sample sizes of 1,000 interviews, and even better considering the effective sample size of around 600, would mean a margin of error of, say, +/-4 per cent.
Of course the acid test will be on May 5. But if the past is any guide, the easy critique made by both naïve interviewers and pundits who should know better, should be reconsidered, and their cacophony reduced. At least until May 6, when we'll all know if by applying this we've introduced some other source of error.
But then if it was an easy business everybody would be in it, instead of being our self-appointed critics without the responsibility of meeting unreasonable expectations themselves.
|
