The pollsters got it wrong – but they are still crucial for politicians
Published by The Daily Telegraph (20th January, 2016)
Shortly before the 2010 general election, David Cameron sat down with the results of the latest in-depth opinion poll. It was a challenging time for the Tory leader, who had just a small lead over an unpopular Labour prime minister and a vote looming that would determine his future. His personal ratings were holding up, but he knew there were plenty of enemies in his party and beyond waiting for him to fail.
Despite the intense pressure, Mr Cameron spent almost an hour ploughing through the findings. One observer watched him circle key points with his pencil and ask a barrage of questions. At the end, the politician fighting to be prime minister looked up and asked: ‘Why do people say we should not take an interest in the polls?’
Cameron has never pretended being anything other than extremely interested in what people think of him, his policies and his opponents. Last year, however, confidence in the dark arts of opinion polling was shaken by the general election, when survey after survey predicted the race too close to call. Almost two-thirds of polls published before voting day last year indicated either a dead heat or Labour victory, distorting the campaign and influencing the result. The Tories won by 6.5 per cent.
It is absurdly obvious in retrospect the polls were wrong. Ed Miliband was not a credible prime minister, the Tories were more trusted on the economy and Mr Cameron more popular than his party. I curse myself for saying privately the Tories would win, then falling for poll-influenced conventional wisdom in public. Now we have an equally obvious conclusion to the polling industry’s long-awaited postmortem into their catastrophe: they failed to reach enough Conservative voters.
The inquiry found two key problems. First, there were too few over-70’s in pollster samples. YouGov accepts it mistakenly bundled up its 16,000 respondents in this age sample with those 10 years younger. ‘We made a foolish error,’ said Stephan Shakespeare, the firm’s boss – although in their defence, this probably reflects the changing nature of an aging society that sees growing divergence among a group once lumped together as ‘elderly’. Then all the companies over-represented a load of unusually engaged, Left-leaning younger voters.
This is not the first occasion pollsters have been publicly humiliated. Indeed, it is the third time in under half a century that there have been signs of anti-Tory bias in their general election methodology. Ted Heath won in 1970 despite most surveys suggesting a Labour win, then John Major defied all the polls predicting his defeat with a seven-point victory.
Right or wrong, polls are significant. We saw what happens without them in last year’s Oldham by-election, when Ukip claimed to be on brink of victory before its falsehoods were exposed by Labour’s comfortable win. Polls can also be of crucial importance in shaping campaigns. Both Downing Street and former SNP leader Alex Salmond believe one single poll saved the Union: the YouGov survey published 12 days before the independence referendum that predicted a ‘Yes’ vote. It forced Westminster’s leaders to get their act together just in time.
Yet though they still matter, it is getting ever harder to carry out polls as the two-party system breaks down, society becomes more complex and technology changes. One firm carrying out phone surveys used to take 4,000 calls to garner 2,000 responses; now it needs 30,000, since most people use mobiles and, feeling harassed by cold callers, are less keen to co-operate. Another company claimed it is hard to reach all the correct cohorts of older voters with online questions.
Perhaps the biggest question is whether politicians pay too much attention to polls, since there is a popular belief that – devoid of principles – they switch direction according to the latest survey. ‘I have never met a politician who was not interested in polling – apart from John Redwood,’ one researcher told me. Another expert said Michael Heseltine was so excited by a poll revealing he was more popular with voters than Margaret Thatcher that he rushed into Mori’s office and printed off copies himself.
There is no shame in that. Politicians should take notice of polls. Members of the political class exist in a stifling bubble. It is hard to appreciate the intoxicating and mind-warping nature of Westminster without experiencing the system from inside a party bunker. So we should not begrudge those governing the country any information they can get on how to prioritise issues, communicate their policies and understand public concerns. ‘Without them [polls], a modern leader is driving blind,’ said Jonathan Powell, Tony Blair’s chief of staff.
This is why the current Conservative leadership has conducted monthly surveys delving far deeper into the sentiment of the nation than simplistic queries on voting intention. The Tories want to leave nothing to chance. Their surveys have included up to 100 questions: the regular ones on trust, the economy and perceptions of the party, alongside perhaps 30 grappling with a particular issue and breaking down responses according to age, gender, race and social class. Complex models are built from the responses: one influential analysis plotting ‘security’ (essentially economic) alongside ‘diversity’ explains Labour success in London and rings alarm bells for the future.
As the polling sector seeks to restore its shattered image after last year’s disaster, it faces another challenge with the European Union referendum. This presents a paradox for the Prime Minister. On one hand, he would like to see a hefty majority against Brexit to ensure any wavering cabinet colleagues with one eye on his job remain loyal to his position. But polls predicting this result might encourage people to vote Out who would not otherwise do so – people who would love to give Brussels a kick yet are nervous of actually leaving the EU.
In usual times what would suit Mr Cameron best would be a series of polls consistently predicting a narrow victory for ‘In’. That would keep everyone in line and on message. But these are not normal times. Just ask the pollsters whose last confident, repeated forecast of a narrow victory at the polls would have seen Ed Miliband in No 10.