The referendum could make or break reputations
Just a week to go before the big day and there’s one group that’s getting very nervous – the pollsters. This is such a massive election that their final polls will be remembered for years just as now we point to surveys in Quebec in October 1995.
The main challenge is that they’ve had no experience of surveying opinion for an election like this and there are so many uncertainties.
Take turnout for instance. It is one thing to tell a pollster that you are 10/10 certain to vote and another to actually go to the polling station to put your X on a piece of paper.
Will those saying they’ll vote who have never voted before actually do do? The YouTube clip above shows the efforts the authorities are making to take the mystery out of the process.
We know from previous elections that the most likely voters are those who have done so before which is why ICM down-weights by 50% those who did not vote in the previous election. Will that ratio be valid this time?
At the 2011 Holyrood election, the one that the SNP won and which triggered off the referendum, had a turnout of just 50%. That’s a much lower figure than the 80% being suggested for September 18th. At the 2010 general election fewer voted in Scotland, 63.8% than in the UK as a whole, 65.1%.
We hear a lot about the “shy Noes” – those who are not convinced but reluctant to tell a pollster their true view. We had this in the 90s over “shy Tories” which was said to be one of the reasons why the industry as a whole got GE1992 so wrong.
How are efforts to deal with that going to work out. So many imponderables so much at stake.