We are now (obviously) in the final phase of the campaign – and thus into the period when, if it is going to happen, the process whereby voters switch back to the status quo (i.e. Remain) because of concern about the risks of change might now be expected to be in evidence. Indeed, the polls published at the weekend that showed some recovery in Remain’s position after a set of poor polls (from their perspective) last week, could be interpreted as providing such evidence. The latest polls, however, do not.
All in all, since the weekend three new polls have been published in full, one each from ORB, Survation and YouGov, while a further poll from Survey Monkey has also been reported but no technical details seemingly been made available (which means at present at least the figures for this poll are not reported elsewhere on this site). (Survey Monkey are not members of the British Polling Council and are thus not subject to its rules on transparency; this is not the first instance of a US-based pollster reporting findings about the UK’s referendum without publishing any technical details, even when they have been requested.) Three of these polls were conducted after the tragic murder of Jo Cox, while in the fourth case (ORB) interviewing took place both before and after. In two cases (Survation and YouGov) their latest reading is the second poll that they have conducted since that sad event.
If voters are now swinging back to Remain in the final days of the campaign, we might expect some of these polls at least to be showing an increase in support for Remain. In practice none of them do so. YouGov’s latest offering, conducted between Friday and Sunday, puts Remain on 49%, Leave on 51% (after Don’t Knows are excluded); that represents a two point swing to Leave since the company’s previous poll, conducted on Thursday and Friday. Survation, in a poll conducted on Monday, put Remain and Leave neck and neck on 50% each (again after Don’t Knows are excluded); this also represents a two-point swing to Leave since their previous poll conducted on Friday and Saturday. In short both the companies that were polling for a second time since the murder of Jo Cox have reported a slight swing back to Leave.
Meanwhile, Survey Monkey, in a poll was apparently conducted between Friday and Monday, put Leave one point ahead (by 49% to 48% with Don’t Knows not excluded), after having put the two sides neck and neck in a previous poll undertaken in the week ending last Wednesday.
The final poll in the quartet, from ORB, might in contrast be thought to be providing evidence of a swing to Remain, albeit that much of its fieldwork is relatively old. Its headline figure was Remain 54%, Leave 46%, representing a one point swing to Remain as compared with the headline figure in the company’s previous poll. However, this apparent swing to Remain is the product, firstly, of a decision by the company to take as its headline figure for this final poll the responses of just those who said they were certain to vote together, and, secondly, of an unusual pattern (not only for ORB, but also for nearly every other poll conducted during the referendum campaign) whereby those who said they were going to vote to Remain were more likely to say they were certain to vote. (ORB also assumed that those who said they did not know how they would vote would break three to one in favour of Remain, though this imputation only seems to have involved nine respondents.) If this poll had been headlined in the same way as previous ORB polls had been (at least by the company itself), the figures would have been (after Don’t Knows are excluded) Remain 52%, Leave 48%, or a one-point swing to Leave.
Now, of course, in all four cases the reported swing to Leave is small, just one or two points, and could well simply have been occasioned by the chance variation to which all polls are subject. We certainly cannot presume that any such movement to Leave has actually occurred. However, what is clear is that there is no evidence at all, in these four polls at least, that a gradual, continuous swing back to Remain is now taking place as voters weight up the relative risks of the two options before them. All that they do suggest – as did the weekend’s polls – is that the position of the Remain side may not be quite as weak as it appeared to be at one stage last week.
Not that the Leave campaign has been particularly effective at persuading people that exiting the EU would not be a risk. YouGov report that 53% now think that leaving would be risky, more than have done so on any previous occasion that the question has been asked, while only 34% believe it would be safe. Meanwhile, ORB report that 56% associated the phrase ‘is a risk’ with the Leave campaign, while only 27% regard the Remain campaign in the same way, figures little different from those the company has reported before. So the potential for a swing back amongst voters who are concerned about risk may well be there – so far, however it is not clear that it is actually being realised.
Consequently, the referendum still looks very tight. On average the last six internet polls (excluding Survey Monkey) have put Remain on 49%, Leave on 51%, while the last six phone polls reckon that Remain are on 51%, Leave on 49%. The difference between the two kinds of poll is now seemingly much narrower than it has been for most of the campaign, a convergence that seems to have consisted primarily of phone polls moving closer to the estimates provided by those conducted over the internet. Meanwhile, if as NatCen’s research suggests might be the case, the eventual result could well fall somewhere in between the estimates of the two kinds of polling then we cannot be sure which side is ahead. There seemingly really is everything to play for in the last 24 hours.