There are just two things polling organisations have to do when it comes to general elections. They must be able to establish a representative sample of the electorate and they must have a clear idea of the reliability of what those people say about their voting intentions.
Unfortunately our pollsters are unable to do either very well. Typically, this has resulted in a marked underestimation of the Conservative vote. It has been excruciatingly embarrassing when the battle between Labour and Conservative is close, which in 1992 and 2015 had the pollsters fooling people into imagining there would soon be a Labour prime minister in No 10. Alas for Neil Kinnock and Ed Miliband, until the results piled up neither had seen the need to prepare a concession speech. Through little fault of their own, they were made to look the biggest fools of all.
So at the outset of the current campaign the polling organisations might have felt relieved at such an apparent gulf in public support between the Conservatives and Labour. Such firms’ reputations aren’t on the line when the race is not close.
“This race isn’t close. But that is not the picture being painted by some notable recent polls
Having talked to candidates of both main parties on the field of combat, I still don’t think this one is, remotely. But that is not the picture being painted by some notable recent polls. Last week Ipsos Mori had Labour smashing through the 40% barrier and YouGov produced analysis suggesting the Tories were facing a loss of seats to Labour, leading to no overall majority in the House of Commons. This would be a near-terminal rebuff to a prime minister who had gamed the Fixed-Term Parliaments Act to hold an election of choice — especially as Mrs May has run a clunkily personal campaign (a cult of no personality, as this column put it three weeks ago).
Unlike the 2015 election, when the pollsters were “herding”, this time they have been much more divergent. The reason is that they accepted their failure in 2015 was systemic, not the result of any “late swing to the Tories” — and each of them has been modifying its methodology in a different way. They call it science; I call it intelligent guesswork.
Advertisement
They all have a different solution to the turnout “problem” — specifically, of younger people (18 to 30-year-olds) whose propensity to vote has been much less than they claim when interviewed by the pollsters. Thus, those polls that show the most rapid rise in support for Labour are exactly those that also predict a very high turnout of younger voters. The point is that they, in sharp contrast to the rest of the population, are much more attracted to Jeremy Corbyn than to Theresa May.
But some of those polls showing the highest figures for Labour suggest a turnout on Thursday in excess of 80% on the part of younger voters in their surveys. This may reflect what the young people are saying. But if they were as good as their word, it would be a dramatic break with all recent elections when this demographic has been strikingly less disposed to vote than (predominantly Conservative) older voters.
It was this that confounded the pollsters in 2015: the Milifandom of younger voters was insufficient to mobilise sufficient numbers of them to enter the polling stations. For Corbyn’s fans, there was great excitement that about 110,000 more young people registered on the cutoff date for the current election than did on the equivalent day in 2015. But even assuming that all those actually vote (which they won’t) and that they all vote Labour (which they won’t), this would represent only about 175 extra votes per constituency.
It’s worth listening to someone who has great experience of getting such campaigns wrong: David Cameron’s pollster during the referendum, Andrew Cooper. On the day of the vote he reassured the then prime minister that his firm’s figures showed “remain” a full 10 points ahead; it was not even close.
Last week Cooper observed that — as in the current general election — there had been a dramatic difference between the views of older and younger voters; the former were by a wide margin pro-Brexit, the latter (to a dramatic extent) pro-remain. “When we looked at the demographic profile of the people who had applied to go on to the electoral register in order to vote, it was very heavily skewed towards younger people. The incidence of 18-34s among those new registrants was more than twice what it is among the population. That gave us reassurance that those people really were going to do what they said they were going to do. And what actually happened is that a lot of young people joined the register and then didn’t vote.” The bastards.
Advertisement
There are those who have had great success in reading polls successfully, such as Matt Singh, founder of Number Cruncher Politics, whom (to my benefit) I consulted before the 2015 election. Back then Singh had observed that polls showing Labour in the lead were plainly suspect. He noted that the Tories were well ahead in polls measuring the public’s view of the parties’ “economic competence” and also on which had the better leader. And that no party had ever done well at the ballot box when well behind on these two measures. Singh also pointed out that results in recent local elections had shown Labour doing badly (and they have done even worse in 2017). These real results have always been a reliable pointer to the succeeding general election.
Despite the pronounced narrowing in the overall “voting intention” figures, all the polls show a large gap between the two main parties on economic competence and leadership — in the Tories’ favour. Moreover, in this unusual election, Brexit registers as the most salient issue. On this too, the Conservatives are seen — by a wide margin — as the more trusted.
This is where “Theresa May’s team” is likely to cause most damage to Labour: among electoral tribes who would never before have considered voting Tory. It is in the Brexit-voting Midlands and north where there are prospects of gaining seats that had been Labour “givens” since 1945. This was the message of the Copeland by-election in February. It also explains the unusual nature of the Conservatives’ campaign, concentrating their efforts on apparent Labour strongholds whose electorates voted heavily for Brexit.
The point of a first-past-the-post parliamentary election is not to gain a particular percentage of the overall vote — but to seize the maximum number of seats. Corbyn may even succeed in his real objective of beating the 30% Miliband achieved for Labour — but in an old-fashioned two-party contest that won’t prevent the Conservatives winning their biggest parliamentary majority since Margaret Thatcher’s in 1987. The Tories had a serious wobble in that campaign, too.