A Question of Wording? Another Look at Polling on a Second Referendum

The question of whether or not there should be a second referendum has been one of the hottest topics in the Brexit debate during the summer. In part, the debate has been stimulated by the relatively adverse reaction with which the Chequers Agreement was greeted, a reaction that led some, such as the former Conservative minister, Justine Greening, to advocate the idea of a three-way ballot between no deal, Chequers and remaining. That adverse reaction certainly helped fuel speculation more generally about the difficulties seemingly facing the government, first, in reaching an agreement with the EU and, second, in securing parliamentary approval for the outcome of the negotiations – and the possibility that another referendum might play a role in resolving any resulting impasse. At the same time, the discussion has been stimulated by the efforts of the high profile pro-second referendum Best for Britain and People’s Vote campaigns, who have managed to secure considerable publicity during the relative quietude of the parliamentary recess, and who are attempting to persuade the Labour Party in particular to support a second ballot.

As part of their campaigning, those two organisations have been keen to give the impression that not only is there majority support for a second referendum, but also that this support has been growing. There has certainly been plenty of polling on the subject in recent months, including not least polling commissioned by these campaigning organisations themselves. But how much support is there now for a second referendum, and is there any sign that it has grown, not least perhaps in the wake of the Chequers Agreement?

In our analysis of previous polling on attitudes towards holding a second referendum, we have noted two key features. First, the wording of the question matters; voters are less likely to endorse the idea if asked whether there should be another ‘referendum’ than if they are asked whether the ‘public’ should be able to vote. Second, the form of the referendum matters; Leave voters are, unsurprisingly, rather warmer to the idea of a referendum in which the choice is between whatever deal the government has negotiated and no deal than they are towards one where the choice is between a deal and remaining in the EU.

Meanwhile, one striking feature of the most recent recent polling on the issue is that many very differently worded questions have been asked. Leaving aside questions about a multi-option ballot (on which see here and here), all in all, no less than eleven differently worded questions about a second referendum have been included on published polls since the beginning of April. And as we might anticipate these have elicited rather different patterns of response, from which proponents and opponents of the idea have been able to cherry pick as they see fit. Meanwhile, many of the questions have only been included on one poll, which means that ascertaining whether or not attitudes have changed over time is, despite the plethora of polling on the subject, rather more difficult than might have been hoped.

The table at the bottom of this blog details the variety of polling that has taken place. It shows, first, the exact wording of each question that has been asked, the most recent level of support and opposition recorded in response to that question among voters as a whole, and, then, the level of support and opposition separately for those who voted Remain in 2016 and those who backed Leave. The questions are listed in order of the extent to which voters as a whole expressed support for rather than opposition to the idea. Meanwhile on the right-hand side of the table we identify some of the key features of the wording of each question, and provide brief details of the polling company and the dates of the fieldwork.

One point immediately stands out. Most of these questions have found more people in favour of a second referendum than opposed. But the balance of opinion has varied considerably. In those polls listed at the top of the table twice as many people expressed support for a second referendum as said they were against. In those at the bottom of the table, opponents outnumbered proponents by ten points. In between are a number of polls that exhibit a modest excess of supporters over opponents, while in most instances, although more numerous than opponents, the proportion actually expressing support is rather less than half. (Invariably, many respondents to these polls choose the mid-point ‘neither’ option, if available, or else say that they don’t know.)

We can, however, make sense of much of the variation by reminding ourselves of the two key lessons of previous polling on this subject. First, only three of the questions refer to holding another referendum as opposed to some kind of vote of the ‘public’. These three polls account for three of the four entries at the bottom of our table. Second, only four polls specify what the options would be on the ballot paper, in each case indicating that the choice would be between Leave and Remain. Two of these polls appear at the bottom of our table, while the other two have among the lowest levels of support among those polls that refer to a vote by the ‘public’.

In short, polls that ask people whether the ‘public’ should have a vote usually record a higher level of support than those that just ask whether there should be another referendum. Equally, polls that do not specify what the alternative would be to endorsing whatever deal is reached tend to secure higher levels of support than those that indicate that the choice would be between leaving (on whatever terms have been agreed) and remaining.

But why do these patterns arise? We secure a valuable clue if we look separately at the responses of those who voted Remain in 2016 and those who supported Leave. This reveals a striking contrast. The balance of response among Remain voters is much the same irrespective of the question asked, while there is little or no discernible pattern to the variation that does exists (though perhaps we might not be surprised that support for another referendum is particularly high among Remain supporters when respondents are asked what should happen in the event that there is no deal at all). It looks as though we can say with a considerable degree of confidence that around two-thirds of Remain voters are in favour of another ballot while only around one in six are opposed. Most likely, many of the two-thirds of Remain supporters that respond positively to polling questions on the subject simply assume that remaining would be one of the options on the ballot, irrespective of the precise wording that is used.

The picture among Leave voters, however, is very different. They are more likely to back another ballot when it is presented as a vote by the ‘public’ than as a ‘referendum’ and less likely to support the idea if it is made clear that the choice would be between remaining and leaving. The latter point is, perhaps, not surprising. After all, most Leave voters do not think that the original majority vote to Leave should now be questioned or overturned. The former pattern, most likely, reflects the populist outlook of some Leave supporters that means that they warm to the idea of power being placed in the hands of the people rather than an elite. Given this outlook, it is perhaps not surprising that the one and only question where more Leave voters expressed support than opposition was one where respondents were asked whether, in the event of there being no deal, the decision about what to do should be made by MPs or the public. After all, as other recent polling for The People’s Vote has shown, few Leave voters trust MPs to make the right decisions about Brexit.

So, whether or not there is majority support (or more accurately, a plurality of support) for a second referendum is less clear than might be imagined from an initial, quick glance at the headline polling results. Certainly, if we are to use polling to evaluate the level of support for the kind of referendum being promoted by The People’s Vote, that is, one in which the alternative to accepting whatever deal had been agreed would be to remain in the EU, it would seem essential to refer only to those polls that make clear that the choice in another ballot would be between Leave and Remain. Meanwhile, it is at least debatable as to whether it is wise to use wording that might be thought to be playing into the populist sentiment that exists among some Leave voters. Maybe if the issue does become more pressing in the autumn, opponents of Brexit would prove able to invoke such sentiment in support of another ballot. But equally, perhaps, that support might evaporate if it became clear to Leave voters that a second ballot would revisit the issue of whether Britain should leave the EU in the first place.

Nevertheless, that still leaves the question of whether support for holding a second referendum has increased, and perhaps especially so in the wake of the Chequers Agreement. Until recently, two companies, Opinium and YouGov, had been asking the same question about the issue on a reasonably regular basis. Both their questions referred to a second ballot as a ‘referendum’, and in one instance, Opinium, also made it clear that the choice would be between remaining in the EU or leaving. Both have thus hitherto found more people were opposed to a second referendum than were in favour, though both also suggested that the level of opposition had fallen to some extent at least during the previous year or so.  However, unfortunately, Opinium have not asked their question since the beginning of June, which means that we have to rely heavily on a single source, YouGov, in coming to a judgement as to whether there has been a further significant shift more recently.

YouGov’s data certainly suggest that there has been at least a discernible, if modest shift. In three readings of its regular question that the company took between April and June, it found on average that while 38% were favour of the idea of another referendum, 45% were against. But in three readings taken since the Chequers Agreement, 40% have said they are in favour, almost matching the 41% who are against. Indeed, in its most recent poll that included their usual question, conducted at the end of July, YouGov found for the first time slightly more (42%) saying they were in favour than stating that they were against (40%).

But is this apparent trend corroborated by any other evidence? There is some. Just before the Chequers Agreement was reached, BMG found 44% in favour and only 27% against ‘a referendum being held asking the public whether they accept or reject the terms of the deal’. When they asked the same question again at the beginning of August, 48% said they were in favour and only 24% opposed. On the other hand, the trend has not been replicated when Opinium have asked a different question from the one they had previously been asking. This question asked respondents whether they ‘support or oppose the public having a vote on the final deal that the Government agrees with the EU’. When they asked this on two occasions in April and May Opinium found no less than 53% were in favour and only 31% opposed. But when they asked the question again in July, after the Chequers Agreement, support stood at 50% and opposition at 35%.

So, while there is some evidence that Chequers may have persuaded a few more voters of the merits of holding another ballot, we probably need more instances of the same question showing an increase in support since before the beginning of July before we can be sure that this is indeed what has happened. But in the meantime, we certainly need to remember that this is a topic on which, above all, question wording matters, and where, so far, the wording has perhaps not always conveyed clearly to respondents exactly what kind of referendum is being suggested – or advocated.

By John Curtice

John Curtice is Senior Research Fellow at NatCen and at 'UK in a Changing Europe', Professor of Politics at Strathclyde University, and Chief Commentator on the What UK Thinks: EU website.