Not for the first time in recent weeks, a poll has made waves amongst commentators and the markets. The latest example came from ORB in an internet poll that appeared in Saturday’s Independent. Its headline figures were Leave 55%, Remain 45%. Although the Leave tally in the poll was boosted by two points as a result of a decision to focus on the figures after the data had been weighted by people’s reported propensity to vote (rather than, as had been the practice for this poll hitherto, those not weighted in that way), even the figures when the poll was not weighted in that way (Leave 53%, Remain 47%) still represented the biggest lead yet for Leave in any of ORB’s internet polls.
Not for the first time either, however, a poll that made waves proved to be a bit of an outlier. Two internet polls published 24 hours later reiterated the by now very familiar story from polls conducted via the internet that it is difficult to tell which side is ahead. Opinium for The Observer put Remain ahead by 51% to 49%, just as they did last week. YouGov for The Sunday Times said that Leave were ahead by 51% to 49%, a slight turnaround on the company’s previous figures for The Times in the middle of last week.
However, ORB’s poll was not the first during the last week or so to suggest that Leave may be doing better than at any previous stage during the campaign. As we noted in our last blog, ICM’s most recent poll also gave Leave a record lead. At the same time, one of three polls from YouGov last week gave Leave its highest lead since early February. In contrast, no poll has recently reported a particularly good set of figures for Remain. Thus, while the polls have been bouncing around somewhat (which is just what we should expect given the chance variation to which all polls are subject) they appear to be bouncing around a central point that looks a little more favourable to Leave than before.
Certainly it remains the case that on average polls done over the internet since ‘purdah’ kicked in on May 27th put Leave narrowly ahead – by 51% to 49% – whereas previously the sides were both on 50% each. The weekend’s three polls do not change that calculation. There is thus some, albeit still highly uncertain, evidence that there might have been the smallest of swings to Leave during the last fortnight. Meanwhile, of course, we should remember that most phone polls, which are expected to be more numerous this coming week, have so far nearly all been putting Remain ahead.
One feature of the ORB poll that did attract particular attention was its finding that as many as 42% of those who voted Labour in last year’s general election say they would vote to Leave. This, it was suggested, was symptomatic of how the Labour leadership’s apparent lack of enthusiasm for Remain was undermining support for the EU amongst the party’s supporters. However, if the overall figures in any poll are potentially subject to chance variation, that is certainly the case for any particular sub-group of respondents. The latest YouGov poll, in contrast, suggests that only 27% of 2015 Labour voters will vote to Leave, while the average level of support for Leave amongst Labour supporters across the most recent poll conducted by each company that is polling regularly in this referendum stands at 33%. This latter figure has changed little throughout the campaign. Rather than providing evidence of a sudden swing to Leave amongst Labour supporters, recent polls have simply underlined the point that Labour has been struggling throughout the campaign to persuade many of its supporters to back its point of view – and it is far from being the only party to find itself in that position.
Much of the detail of this weekend’s polling focuses on voters’ perceptions of the personalities in the referendum campaign rather than on people’s views about the issues at stake. Nevertheless, there is some potentially illuminating new material in YouGov’s latest poll about how people view the two key issues of the economy and immigration.
First of all, it appears that the Remain side’s attempt to win voters over through warnings of the allegedly dire consequences of leaving the EU are regarded as over the top by many voters. Only 12% say that the Remain side’s claims about the economy are ‘realistic and true’, while as many as 35% say that not only are they exaggerated but that ‘there is very little truth to them’. In contrast as many as 23% think the Leave side’s claims about immigration are true, while only 29% reckon there is little truth to them. At the same time, no less than 41% feel that the Remain campaign has been the most (sic) scaremongering of the two campaigns, while only 28% reckon the Leave campaign has been. It would seem that the Remain side needs to bear in mind that if its warnings appear to be too shrill they may simply not be believed.
Second, YouGov’s poll also reveals a potentially important asymmetry in how voters view the two issues of the economy and immigration. On immigration, voters are not only inclined to believe that it would be higher if we remain in the EU (52% of all voters and 81% of Leave supporters express that view) but also that it would be lower if we left (49% of all voters and 74% of Leave supporters reckon that would be the case). In other words, when it comes to immigration many voters regard leaving the EU as a potential solution to what they view as a significant problem.
On the economy, however, the picture is rather different. While 45% of all voters and as many as 84% of Remain supporters think that leaving the EU would damage the economy, only 19% of all voters and 41% of Remain supporters think that staying would prove positively beneficial. This suggests that while the Remain side may have been relatively successful at persuading voters of the economic disadvantages of leaving the EU, it has largely failed to persuade them of the benefits that would flow from staying. Given that the debate about the economy appears to be the Remain side’s strongest card, this asymmetry of perspective on the issue might be regarded as an important strategic weakness in its case.
Meanwhile, the debate about the importance of ensuring that polls of referendum vote intentions are representative of the population in terms of voters’ educational background continues. While Opinium have not opted to weight their data by educational background (though they are now doing so by how socially liberal or conservative people appear to be, a phenomenon that varies accoding to educational background) they have published details of the proportion of graduates etc. in their latest poll. They report that 37% of their sample have a degree or equivalent professional qualification, a figure that may not be wholly out of line with the 35% who were classified in the most recent British Social Attitudes (BSA) survey as having a degree or other higher educational qualification. However, only 7% of Opinium’s respondents are classified as having no formal qualifications, well below the 18% figure in the most recent BSA. But given that those with fewer qualifications are more likely to be in favour of leaving the EU, any deficit in this respect in Opinium’s sample can hardly be used to help explain why its internet polls, like those of other company, are more likely to be more favourable to Leave.
One company, however, that has now decided to pay more attention to the educational backgrounds of its respondents is Ipsos MORI. In conducting their phone polls they now propose both to use educational background as one of the criteria used in sampling respondents and to weight their final sample by educational background. Their move is an acknowledgment that the claim that phone polls contain too many graduates may in their case at least be correct and that this could help explain why their polls typically report better results for Remain. As a result, the company state that if it had been applied to their last poll, the change (together with checking whether respondents are registered to vote) would have taken two points off the Remain tally. Such a reduction will not be enough to close the gap between internet and phone polls entirely, but thanks to this and a decision by Ipsos MORI now also to take into account who is more or less likely to vote we should not be surprised if the company’s next poll (due this week) gives Remain rather less than the 60% with which it was credited in its last poll. We now await what the coming week will bring.