Looking back, the hubris was extraordinary. Courtesy of cut-price internet surveys and the deep pockets of Lord Ashcroft, general election 2015 was awash with more data than any before. The abundant numbers were crunched into unwisely precise predictions about the next House of Commons. The political class – yes, including the journalists – forgot that the quality of information counts for more than the quantity. The failure to spot David Cameron’s majority coming embarrassed players from across the field, and for those on the Labour side obscured an urgent need to sharpen the message. Britain’s politics was left blinded by semi-science.
The trust the pollsters lost may never be completely restored. That’s no bad thing; claims to see the future should always provoke a little scepticism. And yet data-free politics is even more prone to confusion than politics rooted in flawed polling. Ahead of the Oldham West byelection last month there was another surge of groupthink about Labour’s chances based less on statistics than lies. And remember this: a few months earlier, Britain awoke to the barely conceivable reality of the career rebel and initial long shot Jeremy Corbyn leading Her Majesty’s opposition only because a couple of bespoke polls showed him building an unstoppable lead. As rage about the big miss last year recedes, the polls’ success in predicting several other earthquakes – the SNP landslide in Scotland, the collapsing Liberal Democrat vote and the extent of the Ukip surge – look more impressive.
For people who care about politics, flawed polls remain better than total ignorance. For Labour MPs trying to decide how much slack to cut Jeremy Corbyn, it is better to be guided by imperfect and often dispiriting surveys than blind prejudice. Meanwhile, as polls give out alarmingly mixed messages about the looming EU referendum, anybody tasked with running an effective campaign on an almighty national decision would surely like to see the ambiguity cleared up.
Tuesday’s Sturgis review into what went wrong last year is thus very much to be welcomed. Drawing firm conclusions from flawed data is not easy, but the inquiry has moved an impressive distance. Lurking in the dregs of May’s polls it found scraps of evidence against several obvious theories, such as lazy Labour supporters failing to turn out, or a late swing to the Tories, and the pollsters’ failure to speak to the right people emerges as the last explanation standing. A couple of large face-to-face surveys, more rigorously constructed than regular polls, have then been tapped to pinpoint the voters they missed – the very old, the apathetic young, and the “too busy to talk”. The inquiry will now move from diagnosis to prescription. Anybody who wants to see politics rooted in sound information about what citizens think should wish it well.
