Although local and devolved elections taking place on May 5th might have been expected to compete for the limelight, the EU referendum is proving to be remarkably intensively polled. All in all, no less than 21 published polls of referendum vote intention were conducted wholly or partly in March, or more than one every other day. Not that all of this polling is being paid for – quite a few companies are issuing poll results for which there is no immediate client. But if anyone thought that the difficulties that beset the polls in last year’s general election would result in a relatively poll-free referendum campaign, they should by now be disabused of that expectation (or hope).
A Close Race
One probable explanation for the intensity of the polling is that (if the polls are to be believed) the race appears to be close. While most polls put Remain ahead, few do so by much. Of those 21 polls conducted in March, no less than 17 put Remain ahead. But only two gave Remain more than 55% of the vote (once Don’t Knows are left to one side).
Polls conducted over the internet have, of course, long been suggesting that the contest is tight. Five companies (BMG, ICM, ORB, TNS and YouGov) that poll in that way have during the last fortnight or so issued findings that we can compare with equivalent polls conducted a month or so earlier (that is, in late February or early March). A month ago these five companies on average scored the contest, Remain 50, Leave 50 – and the figures are just the same in their most recent polls. ‘Very tight, nothing changing,’ is their persistent message.
In contrast, polls conducted by phone have hitherto tended to paint a relatively optimistic picture for Remain. That, however, has come to be less the case. Three companies (ComRes, Ipsos MORI and Survation) have in recent weeks published polls that we can compare with equivalent polls conducted a month earlier. They now on average score the contest Remain 55, Leave 45. While these figures are still more favourable to Remain than those provided by the internet polls, previously these phone polls credited Remain with 59 and Leave with just 41. And while in their most recent phone poll ORB’s figures have come more into line with those of other phone polls (their first phone poll in mid-March actually put Remain slightly behind on 49%), at 54% for Remain, 46% for Leave, it still pointed to a tighter contest than phone polls had previously been suggesting.
Quite why the figures in the phone polls should have perceptibly narrowed when there is no sign of any movement in the balance of opinion in the internet polls is a bit of a mystery, much as is true of the reason for the original discrepancy in the first place (despite Populus’ valuable attempt to understand what might be happening, on which we have commented earlier). And, of course, maybe the phone polls will swing back in Remain’s favour again. But in the meantime the narrowing of the Remain lead in the phone polls has certainly helped reinforce the impression that the race really is a rather close one.
In any event, a look underneath the bonnet of the polls helps us understand why the contest is so tight. Poll after poll finds that most voters believe that being in the EU results in too much immigration and undermines Britain’s sovereignty. At the same time, most voters are also inclined to the view that leaving the EU would be harmful to the economy and poses something of a risk (perceptions that the UK government’s controversial leaflet setting out its case for staying in the EU appears intent on reinforcing). Many a voter is seemingly having to work out which set of feelings they should follow. Nowhere is this more the case than amongst Conservative supporters, who (despite the Conservative UK government’s position) remain deeply divided on whether to vote for Leave or Remain, pulled it seems in opposite directions by the respective persuasive powers of David Cameron and Boris Johnson.
The Importance of Turnout
Given the apparent closeness of the contest, speculation has inevitably turned to the possible impact of turnout. So far as the demographics of the race are concerned, it is not immediately obvious that one side should necessarily have an advantage over the other on this front. While Remain is much more reliant than Leave on younger voters, who are less likely to make it to the polls, Leave is more reliant than Remain on less well-off, working class voters, who in general are also less likely to vote. However, perhaps one side’s voters are more committed to their cause irrespective of their age or social background.
For the most part this remains an issue on which many polls are still silent. Some companies (such as BMG, ComRes and YouGov) are not as yet even asking people how likely they are to vote in the referendum (although ComRes are producing figures based on their estimate of the differential propensity of people in different social groups to vote in last year’s general election). Others are asking people how likely they are to vote, but are not using that information in their estimates of Remain and Leave support (and in some cases are not even tabulating the figures by referendum vote intention). Only Survation and (in their last couple of polls) ICM are weighting respondents by their reported propensity to vote in the referendum.
Such evidence as we do have, however, consistently points to Leave voters reporting on being keener to make it to the polls. In four recent polls (from ICM, Ipsos MORI, ORB and Survation) for which the information is available, on average just two-thirds (66%) of Remain supporters said they were certain to vote, compared with three-quarters (75%) of those backing Leave. Of course, voters may not in the end prove to be good predictors of their propensity to make it to the polls. Nevertheless, expect the polling companies to pay more attention to turnout in the coming weeks – a change that could simply make the race look even closer.