In December 2018, The Guardian published an article by Peter Kellner, former YouGov President, which claimed in its headline: ‘support for staying in the EU has rocketed’.

Neither the article nor polling at that time supported this claim.

In short

All polls show Remain ahead: In questions where the referendum question was repeated (or its variants), Remain was been ahead, at about 53%.
Different questions: Some polling companies have asked about a choice between the government’s exit deal and remaining in the European Union.
Caution over assumed voting intention: Kellner’s article also drew attention to voting intention under hypothetical scenarios. We should be cautious.

Remain ahead

The main claim of Peter Kellner’s article is:

For most of this year, polls have shown remain ahead of leave, typically by four to six points. But in a referendum between staying in the EU and leaving on the terms that the government has negotiated, staying enjoys an 18-point lead: 59–41%.

It is clear in the article — but not in the headline — that this is to compare two different types of questions.

The What UK Thinks EU website calculates a rolling average of three wordings about the principle of remaining or leaving the European Union:

At the end of December 2018, the polling average (with those who say they don’t know or won’t vote removed) was 53% to Remain:

There has been no rocketing support for Remain. (Source: What UK Thinks EU)

The Remain share has not rocketed upwards, but it has gently risen like a hot air balloon.

Kellner correctly states the typical average lead is four to six points for Remain. The last poll to show Leave ahead was in March 2018, conducted by ORB (and sponsored by the Sunday Telegraph).

This is the wording for the other type of question Kellner refers to:

If there was a referendum tomorrow with the option of accepting the government’s Brexit agreement or remaining in the EU, which would you support?

Since this agreement has been recently published, there is limited polling on this choice. By the end of 2018, only seven polls were published that directly asked this question by British Polling Council members. It has been asked three times by YouGov:

There has been no major rise in support for the EU under this wording. (Source: What UK Thinks EU)

If this is a rocket, then its launch has faltered — with similar readings for Remain in all three YouGov polls (45% — 47%).

As we can observe, there is still a high proportion of people who they do not know how they would vote in such a referendum.

Two additional polls — by Deltapoll and YouGov — have asked people to order their preferences between exiting under the government’s deal, leaving with no deal, and remaining in the EU. These aggregated preferences then yield implied referendum results for pairs of different options (such as exiting under the government’s deal versus remaining in the EU).

Some Leave voters are undecided as to how they would vote in such a referendum, believing the UK should leave in a different manner. A hypothetical referendum turning into a real nationwide vote may settle some minds.

“Imagine that, at the next general election…”

Kellner’s article then draws our attention to voting intention questions asked under hypothetical scenarios:

YouGov asked people how they would vote if Labour, along with the Conservatives, supported going ahead with Brexit. Labour slumps to third place, with 22%, behind the Liberal Democrats, who would jump to 26%.

YouGov surveyed 5,043 GB adults via their internet panel on behalf of the People’s Vote campaign, between 12th and 14th December 2018.

Kellner is comparing voting intention under two hypothetical scenarios (I give the wording for English respondents here, with added emphasis):

That is a large shift. (Source: YouGov)

This shows a notable shift, with Labour’s vote share reducing by 14 points.

These figures exclude those who give non-affirmative responses. If we include people who say they would not vote, they don’t know or refuse to answer, the reduction in the Labour share is 11 points (and the rise in the Liberal Democrat share is 10 points).

If these conditions are realised, will that many voters actually change their support? People may not understand their own vote intention that well, and may not deal with these hypothetical scenarios. The question itself gives prominence to exiting the EU, which is only one part of why people choose to vote. As Anthony Wells of YouGov has shown, polling on new parties and candidates typically provides a large overestimation of eventual performance.

Some voters may be unhappy about the Labour party’s stance, but whether they would choose to support another party is a different matter. We should be cautious about drawing strong conclusions from such questions.

This blog looks at the use of statistics in Britain and beyond. It is written by RSS Statistical Ambassador and Chartered Statistician @anthonybmasters.