Scotland’s independent think tank
Scotland’s independent think tank

The real question a ballot paper on Scottish independence should ask – Paul Jourdan

If a second referendum is to be held on whether Scotland should leave the UK, what should the question be, and why does it matter?

Readers will know that a referendum was held in Scotland on 18 September 2014 in which the question put to the voters was “Should Scotland be an independent country?”, to which the possible responses were “Yes” or “No”, in that order.  Turnout was high at 84.6%, and “Yes” gained 44.7% of the vote, while “No” won the day with 55.3%. The event was known widely as the indyref. Many readers will by now also have seen that in the draft Referendum Bill published on Monday (22nd March 2021) the Scottish Government has made it known that it wishes to ask exactly the same question as in 2014, with the same choice of answers. This creates some urgency in considering whether this is an appropriate and fair choice of wording.

What not all readers will know is that on 28 May 2019 the Scottish Government introduced the Referendums (Scotland) Bill, an Act which was duly passed into Scots Law on 19 December 2019, and that this piece of legislation, as it was initially drafted, attempted to ensure that a second referendum, so much desired by Scotland’s ruling party, would be able to use the same question and answers as the 2014 referendum without any referral being made to the Electoral Commission. Much of legal commentary on the Bill at the time[1] concerned itself with the way in which it aimed to establish a framework for the Scottish Parliament to be able to call referendums on any question within its powers. The idea was that a legal challenge could then be mounted at a later date to determine whether a referendum on leaving the UK should be a reserved power belonging to Westminster, requiring a so-called section 30 order (as was granted in 2013 for the 2014 referendum), or whether the Scottish Parliament can act unilaterally in this regard if it is unable to get the requisite agreement from Westminster. Without the Referendums (Scotland) Act, acting unilaterally would be impossible. In addition, the Bill originally sought to allow the Scottish Government to call a referendum via secondary legislation only, without requiring any further votes to be taken in the Scottish Parliament on the matter.

However, the main focus of criticism from all of the opposition parties was on the attempt to avoid the scrutiny of the Electoral Commission over what the question should be, if a re-run of the 2014 referendum is facilitated by the Act (something which they believe is the sole purpose of the Act). The National reported on 19 December 2019 that “Tory constitutional relations spokesman Adam Tomkins and Labour constitution spokesman Alex Rowley both argued the provisions in the Bill around the questions used in referendums were an attempt to ‘rig’ any future vote by the SNP”, later noting that Liberal Democrat MSP Mike Rumbles was of the same view.[2]   

Given the technical nature of the drafting of the Bill, it is best to explain its meaning through the summary given by the Scottish Parliament Information Centre (SPICe) briefing note:

“The result is that where a question has previously been considered by the Electoral Commission, the Commission’s view has neither to be sought again, nor re-communicated to Parliament at the time of laying regulations [by secondary legislation] or, as the case may be, after the introduction of a Bill. There is no similar provision to Subsection 7 in PPERA [the Political Parties, Elections and Referendums Act 2000 which governs the conduct of referendums in the UK], meaning that on each occasion [in the UK] the Electoral Commission is required to report on the referendum question.” [3]

What was the Scottish Government thinking of when drafting this Bill? How could keeping the Electoral Commission away from scrutinising a referendum question possibly be compatible with the assertion by Mike Russell, Cabinet Secretary for Government Business and Constitutional Relations, that “The rules [the Bill] sets out are of the highest standards and will ensure that the results are widely and internationally accepted”[4]? In the context of the unwritten UK constitution, which has at its core the principle of Parliamentary sovereignty, referendums are problematic at the best of times, seeking as they do to bypass or potentially overrule this very thing. The Brexit referendum revealed problems of this sort in abundance, leading to legal disputes at the Supreme Court and causing years of confusion in the aftermath about what exactly the consequences of the referendum should be, given a situation where there was inadequate support in Parliament in favour of it. The best one can say is that this was an opportunistic use of the SNPs domination of the Scottish Parliament to set up a new referendum on departing the UK by fair means or foul. 

However, in drafting, any legislation concerning referendums needs great care to be taken both to ensure legitimacy and clarity of process.  The Scotland (Referendums) Bill was careless on both fronts, even accepting that its primary purpose is to disrupt the previously accepted consensus on the need for a Section 30 order to be granted by Westminster before a repeat of the 2014 referendum could be held.  It was careless both because it sought to bypass the Scottish Parliament by allowing the referendum to be called by secondary legislation, and because it sought to avoid the scrutiny of the Electoral Commission over the question to be asked. Strangely the Bill at no point considers why it might be that a referendum should be held on the same question twice, or seek to put any kind of limitations on how often the same question can be put. There is no recognition of the danger of allowing a determined government to keep asking the same question until it gets the answer it wants, or of the asymmetry of risk in doing so with constitutional matters, where if a change is assented to it can never be undone, but if it is not, the question can simply be re-asked. Rather, the Bill tries to avoid the bother of having to consult the Electoral Commission all over again when the repetition occurs. There is also no consideration given in the Bill to what kind of new safeguards and processes should be put in place as a result of lessons learned from the shambles of the 2016 referendum, which was characterised by a completely new level of voter manipulation via social media using methods from data science.[5] Furthermore, there is no consideration given to whether a pre-legislative referendum, such as those of 2014 and 2016, should necessarily be followed by a confirmatory vote once the legislation is drawn up, to avoid decisions about permanent and hugely significant constitutional change being made while the public are having to guess about what it will mean in practice.[6] 

Amendments to the Bill were discussed by the Finance and Constitution Committee of the Scottish Parliament on 27 November 2019.[7] The Scottish Government had already backed down on its ambition to be able to call referendums via secondary legislation at this point. Adam Tomkins then brought up the topic about scrutiny of the question by the Electoral Commission “for what are, in essence, repeat referendums”. He cited evidence taken from the Electoral Commission itself: “The Commission firmly recommends that it must be required to provide views and advice to the Scottish Parliament on the wording of any referendum question . . . regardless of whether we have previously published our views on the proposed wording.” He proposed an amendment which would require repeat referendums to be subject to fresh scrutiny by the Electoral Commission. Mike Russell, speaking for the Scottish Government, did not accept this, and instead the Act was passed with a compromise that meant the Electoral Commission would only need to be consulted if the question being repeated was put before the start of the prior Parliament, which at the time the Bill was passed would have allowed the 2014 question to escape a second scrutiny, but not if the referendum is called after the May 2021 elections, which we now know is the earliest that this could happen. 

So if a repeat of that referendum is called in Scotland after the elections this year, while the draft Referendum Bill makes it clear that it wants the question and format to be identical, it is by no means clear that it will get its way. I believe it is unlikely that the Electoral Commission would recommend that the question from 2014 be allowed to run again unamended, and that this is precisely why the Scottish Government went to such lengths to avoid their scrutiny. The reason for this is that there were some deep flaws in the 2014 question which became apparent from the experience of that referendum, and which are more apparent still following the experience of the 2016 Brexit referendum. 

Under PPERA the job of the Electoral Commission (EC) is to comment on the “intelligibility” of referendum questions. In doing so, however, they will comment on their neutrality as well. Their reports make clear that there is an inextricable link between neutrality of the question and legitimacy in a referendum. As well as conducting their own assessment by writing to individuals and organisations, including those representing political parties and campaign groups, the EC also commissions third-party research, which is published in full. This means we can look back at their findings from both the 2014 and 2016 referendums.       

The research for the 2014 referendum was carried out in 2012. The question originally proposed was “Do you agree that Scotland should be an independent country? Yes / No”. This was criticised by the EC as being a leading question in favour of “Yes”, so they recommended removing the phrase “Do you agree that . . . ”, replacing it with “Should . . . ”.  Their full findings were published in 2013.[8] This report mentions one of the principal problems of the proposed referendum: “At the time of undertaking our question assessment, it is not yet clear how a vote in favour of independence would be implemented”[9], adding that in previous recent referendums the question was linked to legislation in Parliament, bringing clarity to what was intended by the question. It would not become clear to the UK public until the debacle following the Brexit Referendum in 2016 what a mess this might lead to, however. In hindsight it appears that the EC gravely underestimated the potential magnitude of this problem in 2012. However, the report noted that the Scottish Government had promised to define what was meant by “independence” by publishing a White Paper with detailed proposals, a commitment that was honoured, and which meant that the 2014 Referendum was held with greater legitimacy than that of 2016, where there was no equivalent of a White Paper or a Prospectus explaining with clarity what leaving the EU might mean. The opening of the report also notes that feedback from their public opinion research was that neutrality would be increased if “factual information about what independence would mean in practice were made available before polling day”, adding that “there was also an expectation that this would happen.”[10] In practice it was deeply naïve to believe that this could happen because almost everything about “what independence might mean in practice” is speculative and subject to intensely divergent opinion. 

IPSOS Mori were hired to conduct the research, and their method was a combination of one-to-one interviews and focus groups with a total of 265 participants in a range of locations across Scotland. In order to test neutrality, they tested three alternative versions of the same question, each with broadly similar wording and the same Yes / No answer. This was not the most rigorous way of doing this, as in all cases they accepted the premise and format of the original question without critique, and without giving their interviewees the opportunity to consider alternatives relating specifically to leaving the UK. The EC would probably argue that this was beyond their remit, however.

The report found that the question was well understood, but that people wanted “more information about what independence would mean in practice”.[11] It found that “Do you agree . . . “ should be removed. It also noted that: “Almost everyone had a clear understanding that ‘independent country’ meant Scotland being separate from the rest of the UK,”[12], although alternative descriptions given initially by people included “separate from England”; “separate from Westminster”; “running our own affairs”; and “Scotland managing on its own”.[13] 

Some participants in 2012 felt that the word “independent” in itself was not neutral, believing that in this context “the word is positively associated with ‘freedom’ and images of ‘Braveheart’ and therefore encourages an emotional and less considered response.”[14] A few participants also suggested “that there could be a potential bias towards ‘yes’ as it comes before ‘no’ “, but recognised that “the same could be true if the answers were reversed.”[15]

Reflecting back on the question posed in 2014 from the vantage point of 2021 the following points come to mind. The most obvious one is that the EC report from 2012 tests whether people it interviewed understood the question by whether or not they realised that the consequence of a “Yes” vote would be for Scotland to leave the UK.  Most did. But this begs a crucial question: if the vote is about leaving the UK, why not make the question about leaving the UK? 

Following the 2016 Brexit referendum certain Brexiteers couched their victory as Independence Day for the UK. Does voting for Scotland to be an independent country mean that it will not only leave the UK but also that it will not join the EU? “Independence” has become an overused and politically laden term. I believe that many more people would now see it as holding some implicit positive or negative implications than they did in 2012. The reason why the EC recommended moving away from a Yes / No format for the Brexit question was because some people they questioned even saw phrases like “remain” or “remain a member” as being potentially positively or negatively charged. How much more so the word “independence” now?

Most people would concede that countries can be independent to varying degrees, rather than this being a simple binary yes or no test. The Collins English Dictionary, for example, gives the meaning of “independent country” as “independent countries and states are not ruled by other countries but have their own government”.[16] Well, Scotland has its own government able to rule on most matters which directly impact voters, such as health, education, taxation, benefit allowance and local government, but is in some respects also governed by Westminster. It has more independence now than it did in 2014 as a result of the Smith Commission and its later implementation. This highlights the absurd ambiguity of the 2014 referendum question, which once noticed can’t be overlooked.  

This can be interrogated further by two thought experiments. Imagine if the Brexit referendum question had been posed as “Should the UK be an independent country?” This would have met with howls of protest from the Remain side, and would have been rejected by the EC as being unintelligible and biased. But is this really so different from asking the same question about Scotland? With the Brexit referendum the Leave side wanted to be free of a union which gave a distant governmental body superior authority to that of the UK Parliament. For them, this was primarily a matter of being independent. 

Second, imagine if the question being put were “Should Scotland be an Independent Country to a significant degree?” Then surely this would be a question that voters in Scotland would say “Yes” to, because Scotland already is an independent country to a significant degree, with its own Government, its own Parliament, its own legal system, its own education system, its own healthcare system and so on.  So the question that the SNP would like to put is very close to one that everybody would agree with, whether or not they wish to leave the UK. This creates a certain pressure to assent, whatever one’s views are. The question is relying on voters to understand a fine distinction between being “independent” and being “independent to a significant degree”. 

These thoughts, in combination with reading the extensive analysis done by the EC in their recommendations about the EU referendum, and the increase in the politically charged nature of the word “independence”, lead to the very firm conclusion that any further referendum on constitutional change in Scotland should adopt almost identical wording to that of the Brexit. This would mean it would read like the following:

“Should Scotland remain in the United Kingdom or leave the United Kingdom?” With the options being “Remain in the UK” and “Leave the UK”, in that order. 

This question would be significantly more neutral because it avoids being coloured by the emotionally charged word “independence” and it avoids any possible Yes / No bias.[17] It would also be much more precise, allowing no room for misinterpretation around what is meant by “independence” and whether the question had implications about EU membership. There is little doubt that with this question any such referendum would achieve significantly greater legitimacy. If some on the Leave side of the argument would object on the grounds that this might make the referendum more difficult to win, then this simply reinforces all of the arguments being made here. In dealing with this matter, the SNP Government will need to decide whether legitimacy matters more than win-at-any-cost expediency. I for one hope that they will reconsider the wording set out in Monday’s draft Referendum Bill, as I believe it is designed to orient a complex debate about the pros and cons of leaving the UK towards the raw sentiments of nationalism, and in so doing chooses to completely ignore the high level of independence that Scotland already possesses. 

Paul Jourdan is CEO of Amati Global Investors

[1] See for example the law firm Brodies LLP commentary at


[3] Sarah Atherton, Iain McIver and Iain Thom, 19 June 2019, p. 14 see–Scotland–Bill

[4] Statement “Next Steps on Scotland’s Future”, 29 May 2019, reported in the SPICe briefing note, p.9. 

[5] This is not conjecture, it was described in detail in Dominic Cumming’s speech to Nudgestock 2017, “Why Leave Won the Referendum”.

[6] The cross party “Report of the Independent Commission on Referendums” published by UCL in 2018 specifically recommended that “where a pre-legislative referendum is absolutely necessary, it should be preceded by a detailed White Paper (recommendation 19) and provision should be made from the outset for a second referendum if plans diverge from those that it sets out (recommendation 20).” (p. 193).  It is reasonable to ask why these serious and important recommendations were ignored in the Bill. 


[8] “Referendum on independence for Scotland: Advice of the Electoral Commission on the proposed referendum question” (Jan 2013). 

[9] Ibid., para 2.3, p.6. 

[10] Ibid., para 2.16 , p.8

[11] Ibid., para  3.18, p. 12

[12] Ibid., para 3.38, p. 15

[13] Ibid., para 3.39, p. 15

[14] Ibid., para 3.45 p. 16

[15] Ibid., para 3.54 p. 18.


[17] The EC reports do not acknowledge any hint of Yes / No bias, but that doesn’t chime with the experience of many in the 2014 referendum.  Some psychologists have written about a strong desire in people to want to say “Yes” rather than “No” to requests for example, see for example “Why ‘Yes’ Is More Powerful Than ‘No’ “, Association For Psychological Science, 4 May 2016.