Michael Gove revealed to be using PR-commissioned puff-polls as "evidence"

Eight out of ten cats prefer Michael Gove to Whiskas.

The Department of Education is notoriously bad at answering freedom of information requests, even being put under special monitoring by the information commissioner's office in December last year because of past inadequacies in answering queries. So it's doubly impressive that Janet Downs, a retired teacher and campaigner who is part of the Local Schools Network, not only managed to get an answer from them, but also extract an excruciating confession about what passes for "evidence" in Michael Gove's department.

Querying a claim made in article in the Mail on Sunday titled "I refuse to surrender to the Marxist teachers hell-bent on destroying our schools: Education Secretary berates 'the new enemies of promise' for opposing his plans", Downs asked for the background to Gove's claim that:

Survey after survey has revealed disturbing historical ignorance, with one teenager in five believing Winston Churchill was a fictional character while 58 per cent think Sherlock Holmes was real.

The department revealed that the main claim sources from a survey "commissioned and conducted by UKTV Gold", and that the other surveys referred to include:

That last survey was linked, by the Department of Education, to an article in the Telegraph, rather than the initial survey.

To be clear, five of the six "surveys" cited by the Department of Education in backing up a claim by a cabinet minister were PR-commissioned puff-polls. They were commissioned, not to find out information in a trustworthy and repeatable manner, but to ensure that stories about UKTV Gold, Premier Inn, the Sea Cadets , Bomber Command Memorial and "teacher-set exam revision service" Education Quizzes found their way into UK papers. Some of them may additionally be respectable polling – the Lord Ashcroft poll around Bomber Memorial Command uses a nationally representative sample, non-leading questions, and face-to-face interviews, for instance – but it's the sort of thing which normally rings alarm bells.

The last cited survey isn't a survey. It's a pamphlet on "Freedom, Aspiration and the New Curriculum" from think-tank Politeia. While it agrees with Gove's conclusion, it is hardly a primary source (an ironic distinction to have to make in a discussion about history teaching).

If this the sort of information which is revealed when the Department of Education responds to freedom of information requests, it's becoming clearer why they so rarely do it.

Photograph: Getty Images

Alex Hern is a technology reporter for the Guardian. He was formerly staff writer at the New Statesman. You should follow Alex on Twitter.

Photo: Getty
Show Hide image

Is Britain about to leave the European Union?

A series of bad polls have pro-Europeans panicked. Are they right?

Is this what Brexit looks like? A batch of polls all show significant movement towards a Leave vote. ORB, a phone pollster, has Leave up four points to 46 per cent, with Remain’s leave cut to four points. ICM’s online poll has Leave up three points, putting Brexit ahead of Remain by 52 per cent to 48 per cent once don’t-knows are excluded. ICM’s phone poll shows Leave up six points, a Brexit lead of three points.

That two phone polls are showing advances for Leave are particularly significant, as telephone polling has tended to show lower figures for Brexit. There is a lively debate over which method, phone or online, is likely to be more effective at predicting the referendum, although no-one knows for certain at the present time.

In any case, whether on the telephone or the Internet, the latest polls have pro-Europeans worried, and Brexiteers jubilant. Who’s right?

There are reasons to start trusting the polls, at least as far as voter ID is concerned

So far, the performances of the political parties in local elections and by-elections has been about par with what we’d expect from the polls. So the chances are good that the measures taken post-2015 election are working.

Bank holidays are always difficult

I would be deeply cautious of reading too much into three polls, all of which have been conducted over the bank holiday weekend, a time when people go out, play with their kids, get wasted or go away for a long weekend. The last set of bank holiday polls gave Ed Miliband’s Labour party  large leads, well outside the average, which tended to show the two parties neck-and-neck.

Although this time they might be more revealing than we expect

One reason why the polls got it wrong in 2015 is they talked to the wrong type of people. The demographic samples were right but they were not properly representative. (Look at it like this – if my poll includes 18 actors who are now earning millions in cinema, I may have a representative figure in terms of the total number of Britain’s millionaires – but their politics are likely to be far to the left of the average British one percenter, unless the actor in question is Tom Conti.)

Across telephone and online, the pollsters talked to people who were too politically-motivated, skewing the result: Ed Miliband’s Labour party did very well among young people for whom Thursday night was a time to watch Question Time and This Week, but less well among young people for whom Thursday is the new Friday.  The polls had too many party members and not enough party animals.

But the question no-one can answer is this: it may be that differential turnout in the European referendum means that a sample of hyper-politicos is actually a better sample than an ordinary poll. Just as the polls erred in 2015 by sampling too many political people, they may be calling the referendum wrong in having too many apolitical people.

These three polls aren’t the scariest for Remain released today

IpsosMori released a poll today, taken 15 days ago and so free from any bank holiday effect, without a referendum voting intention question, but one taking the temperature on which issues the British public believe are the most important of the day.

Far from growing more invested in the question of Britain’s European Union membership as the campaign enters its terminal phase, concern about the European Union has flatlined at 28 per cent – within the margin of error of last month’s IpsosMori survey, which put Britain at 30 per cent. The proportion who believe that it is the biggest single issue facing Britain today also remains static at 16 per cent. Evidence of the high turnout necessary to avert Brexit seems thin on the ground.

Pro-Europeans should be further worried by the identity of the groups that are concerned about the European Union. Conservative voters, the over-65s and people from social grades A (higher managerial, administrative and professional workers) and B (intermediate managerial, administrative and professional workers), are more concerned about the European Union than the national average. The only one of those three groups that is more likely to favour Remain over Leave are ABers, while Conservative voters and the over-65s are likely to vote for Brexit over the status quo.

Among the demographics who are least concerned about the European Union, the only pro-Brexit group that is significantly less concerned about EU membership than the national average are people from social grades D (semi-skilled and unskilled manual workers) to E (state pensioners, casual workers and jobseekers). The other groups that are least concerned with the European Union are people who live in urban areas and people aged from 18 to 24, the two most pro-European demographics.

The prospects of a Brexit vote are rather better than the betting odds would suggest. 

Stephen Bush is special correspondent at the New Statesman. He usually writes about politics.