What to make of the opinion polls? Yesterday, the pound dipped on the strength of a Times splash headlined “Shock poll predicts Tory losses” then bounced back after a Panelbase survey had the Conservatives increasing their lead over Labour to 15 points – quite enough for a landslide. How can reputable professional pollsters be getting such widely divergent results?
Up to Tuesday, the range of poll leads for the Tories in the preceding week was vast – from a fairly modest 5 points to 14 points. Then along comes YouGov in The Times and predicts that Theresa May will lose 20 seats on June 8 and fall short of an outright majority. They cannot all be right.
Broadly speaking, if one goes back to the early days of the campaign – at the beginning of May – the Tory lead over Labour has narrowed, from nearly 20 points to around 10 with the odd outlier pointing to a much closer result. This fits with the consensus that May has underperformed on the stump while Corbyn has confounded expectations of utter incompetence and come across as a more engaging and confident figure. The fact that neither he nor his top team can do simple arithmetic (think yesterday’s car crash interview over the cost of Labour’s childcare plans) has not yet derailed his bid for power.
Andrew Hawkins, boss of the pollsters ComRes, which have tended to give the Conservatives big leads over Labour (12 points at the weekend) has thrown some light on the matter. In an article published on his firm’s website, he explains that the discrepancies mainly come down to how the pollsters adjust for turnout.
We all know that older people tend to favour the Tories and that younger ones are attracted to Labour’s promises of free holidays for all in the Caribbean. The Tories typically lead by 60 points among the over-65s; Labour by a similar margin among the under-25s. We also know that the old are more likely to vote than the young. So Corbyn could be sweeping all before him among the under-30s, but it won’t be of much use to him on June 8 if they mainly stay in bed.
As Hawkins explains,some pollsters, notably ComRes and ICM, adjust on demographic grounds, so the voting intentions of older voters are given greater weight than those of younger ones. Others, notably YouGov, adjust for the self-reported likelihood of a voter going to the polls. So when young people insist that they rock solid certain to vote, they are taken at their word. At present, over 60 per cent of 18-24-year-olds say they are absolutely certain to vote; but only 44 per cent did so in 2015.
This doesn’t quite explain the Times splash, which is not a conventional poll but modelling for a result on a constituency by constituency basis of 50,000 interviews over the course of a week. It also allows for big variations and suggests that the Tories might gain 15 seats on a good night.
So the young have the nation’s destiny in their hands. If they turn out to vote in the same numbers as their grandparents, Mrs May’s much vaunted landslide will turn to dust. But will they defy their history?
(Image: Kevin Walsh)