Lies, damned lies, and (Brexit) statistics

In the heated discussion leading up to the Meaningful Vote originally scheduled for Tuesday the Express reported, from an interview on the BBC’s Andrew Marr show, that the ‘Brexiteer Gisela Stuart masterfully shut down Remain campaigner Gina Miller after she suggested there is now increased support for a second Brexit referendum’.

Gina Miller cited a poll published in the Independent that said ‘People were… for a new referendum by 46 per cent to 30 per cent’, while Gisela Stuart cited research undertaken for Change Britain noting that “The public want their MPs to vote against a second referendum” by 51 per cent to 45 per cent.

Additionally, the Change Britain research also claimed a ‘Canada Plus’ agreement was the most ‘strongly preferred’ outcome while the Independent said a ‘majority of country now think Britain should remain in the EU’. Remarkably, both the Independent and Change Britain polls were carried out by the same company, BMG.

How can this be? How are parliamentarians supposed to make sense of these diametrically opposed conclusions? And, in so doing, how can they make what may be the most important decision of their political careers to guide the country forward for generations to come?

The Northern Ireland peace process was far too important to leave the public opinion research in the hands of partisan organisations and pollsters who would bias the questions, methodologies and analysis to fit their client’s agenda.

In Northern Ireland a programme of independently funded academic polling research was undertaken as a partnership between the Queen’s University of Belfast and the Joseph Rowntree Charitable Trust. Critically, this research was also undertaken with input from all the parties to the negotiations to ensure objectivity so that the results would be taken seriously and acted on.

Clearly, Brexit now needs the same approach. So what has gone wrong with the polling undertaken by BMG and others, and how can it be corrected?

Firstly, with regards to support for a second referendum the result depends not only on when the poll was run but also on the question asked. For example, Lord Ashcroft in his November poll asked ‘Should there be a second referendum, to decide between leaving the EU on terms agreed in the draft Brexit agreement, or remaining in the EU?’, resulting in 38 per cent ‘yes’ and 47 per cent ‘no’ because this question disenfranchises ‘leave’ voters.

Conversely, when asked ‘Should there be a second referendum, to decide between leaving the EU on terms agreed in the draft Brexit agreement, or leaving without a deal?’, the result is only 31 per cent ‘yes’ and 50 per cent ‘no’ because this question disenfranchises remain voters.

However, in a Survation poll also run in November, the result was 48 per cent in support of a ‘People’s vote – a referendum – asking the public their view?’ and only 34 per cent opposed.

In Northern Ireland and around the world people generally like to exercise their franchise, and critically the Survation question does not disenfranchise anyone. Implicitly, both Remainers and Leavers are invited to express ‘their view’.

Interestingly the BMG question gets a result somewhere in-between the Ashcroft and Survation questions, as they ask ‘If there is a vote, should your MP vote FOR or AGAINST another referendum on whether to leave or remain in the EU?’, with an additional option for their MP to abstain.

These very different results now make sense, and the correct approach to dealing with this issue is to either have the stakeholders – the Parliamentarians – collectively agree what is the correct question to ask or, alternatively, run the various alternate questions and then have a discussion as to why they produce different results.

That is the discussion that should have taken place between Gisela Stuart and Gina Miller on the Andrew Marr show, but it didn’t. An opportunity to enlighten the public was lost. Statistics do not have to be lies, they simply have to be misunderstood.

But what about the BMG result in the Independent that suggests the British public want to remain in the EU and their poll for Change Britain that suggests, given a choice, the British public would choose a Canada plus deal? What is happening here?

Regrettably, the polling organisations that have tried to differentiate the British public’s preferences for different Brexit outcomes have not used best practice in both the design and analysis of their questions, pioneered in Northern Ireland and tested in a dozen other countries around the world.

Firstly, the options used in Northern Ireland and elsewhere were drafted by constitutional lawyers who could write both accurate and clear proposals that could be tested against public opinion, while at the same time not leaving any important options out. The eight options tested by BMG do not meet these standards, e.g. No Brexit and No Brexit Plus.

Secondly, BMG used a method that does not allow the informant to separately evaluate every option on offer against all other options by, for example, asking for only a first and last preference.

Similarly, although YouGov use a simple question that asked the informant to rank order just three options, which works well with the alternative vote (AV) system, they then go on to analyse the same data using the Condorcet method that is far from transparent to the average reader.

They would have done better to use the tried and tested methods that worked in Northern Ireland and were published in the Belfast Telegraph for public diplomacy purposes.

The bottom line to all of this is that the British public and MPs are not enlightened by all this public opinion research but rather find themselves frustrated by a lack of clarity, objectivity and transparency. This only leads to further confusion over Brexit in the minds of the British public.

Arguably, the British Parliament has not served the British people well in resolving Brexit. Regrettably, the British public opinion industry has not helped as much as they could in this regard. They could and should have done very much better. Independent research of a higher standard is required.

By Colin Urwin, senior honorary research fellow at the University of Liverpool. This piece originally featured on peace polls.

Disclaimer:
The views expressed in this analysis post are those of the authors and not necessarily those of the UK in a Changing Europe initiative.

View all analyses

Sign up to our newsletter





View our latest newsletter