Posted on 15 September 2014 .
What a difference a week makes. As recently as the end of August, many people in Scotland and the rest of the UK were confidently predicting a No vote in Scotland’s independence referendum. Although their lead no longer looked as impressive as it had in early 2014, No were still comfortably ahead in most polls. Our poll of polls had No at 56% with Yes on 44%.
But then a drip feed of polls started to indicate that things were shifting. A poll from YouGov in early September put Yes on 47% after those yet to make up their mind were excluded. This was the highest tally ever recorded by that company, which up to that point had produced among the lowest estimates of Yes voters. Then the following Sunday, YouGov also became the first company to put Yes ahead – albeit by a statistically insignificant 1 percentage point.
Once the initial Twitter meltdown had subsided, it was clear that what had to some seemed a foregone conclusion was now a race that was far too close to call. Subsequent polls have confirmed this picture. Although, at the time of writing, Yes had only been ahead in one further poll, all the polls published in the last week have put victory for either side at, or very close to, the margin of error - effectively neck and neck.
Our most recent poll of polls puts No at 51% and Yes just 2 points behind, at 49%. In a vote where the winning side needs 50% + 1 of votes cast, it is clear that neither side can now be completely confident of victory.
Which leaves us with the million dollar question – with just three sleeps (or sleepless nights) left to go until Scotland votes, what will swing it one way or the other? In a race as close as this one, it might be considered fairly unwise to be too absolute in any predictions on this front. But nonetheless survey data do offer some clues about the likely factors that may affect the outcome to which we wake up on Friday morning.
One factor that was predicted to affect the outcome, but which is perhaps becoming less significant as polling day approaches, is turnout. Earlier analysis of the Scottish Social Attitudes survey (SSA) by Jan Eichhorn and Lindsay Paterson indicated that those who said they planned to vote Yes were slightly more likely to vote than those who said they would vote No. Accounting for likelihood to vote thus gave Yes a small (but in a tight race, obviously important), 2 percentage point boost. However, predictions of turnout have since been rising. Two of the four polls released on Saturday 14th September (Survation and Opinium) estimate that 90%+ are now 10 out of 10 ‘certain’ to vote on Thursday. And there is now very little difference in voting likelihood between those voting Yes and No. Of course, that finding ought not to prompt complacency in either camp – mobilising their vote will be essential. It is simply that we no longer have a clear basis for saying that turnout will influence the outcome in any particular direction.
What then of the demographic groups that might tip the vote to Yes or No? That women are, overall, less well-disposed towards independence than are men has been well-documented, including in previous blogs by this author. While this remains the case, it does appear that the Yes campaign has made some inroads in narrowing this ‘gender gap’ over time. Using a rolling mean score to even out some of the random variation between polls shows that at the start of 2014, the gender gap in the proportion of men and women planning to vote Yes was typically between 12 and 14 percentage points. More recently, it has fluctuated between around 7 and 10 points (at the time of writing it was 8 percentage points after those who are undecided are excluded). So the gender gap has narrowed but it has not disappeared. Women are also still a little more likely to be undecided than are men, even at this late stage. And as women make up slightly more than half of the electorate, whether or not there is any further narrowing of the gender gap as the remaining undecided women make up their minds may well prove key to the outcome of the vote.
Other ‘dividing lines’ in terms of how key groups look poised to vote are also well known, having been regularly covered by Scottish Social Attitudes. Older people are more likely to vote No – a fact not lost on Yes Scotland, who are planning a last minute mailshot of Scotland’s pensioners to try and persuade them to back a Yes vote. If anything the gap in voting intentions between the oldest age group and their younger counterparts seems to have widened recently, as younger age groups have moved in greater numbers towards voting Yes. For example, TNS’s July poll put Yes votes among the over 65s at 30%, only just below the levels recorded for younger age groups (33-34%). In their August poll, however, the gap was 10-12 percentage points, with over 65s remaining unlikely to vote Yes (31%) while the figures had increased for younger voters to 41-43%.
Those in more working class households are more likely than their middle-class counterparts to back independence. Meanwhile, in terms of party politics, while unsurprisingly most 2011 SNP voters back Yes and most Conservatives back No, Labour voters have become a key target group for Yes Scotland. SSA 2014 showed that Labour voters were among those most likely still to be undecided, while recent polls have indicated that a significant proportion (between a quarter and a third of those who voted Labour in 2011) are now inclined to vote Yes. Better Together campaigners on the left are likely to be targeting as many Labour voters as possible in the next few days to try and persuade them to stick with (or move back to) a No vote.
Meanwhile, SSA has long shown that many people in Scotland hold dual national identities – feeling both Scottish and British to differing degrees. Of these, it is how British people feel that seems to matter the most and, for some, appears to be a reason for remaining part of a political Union with the rest of these islands. In the final days of campaigning, Yes Scotland will try and persuade these voters that independence need not negate any sense of Britishness they feel, while Better Together will counter with arguments about solidarity, common interests and a shared history and culture with people in the rest of the UK.
But with just 3 days to go, and the race on a knife-edge, how much scope is really left to influence voters? Well, all the polls suggest that there are still undecided voters out there. They do, however, vary wildly in how many they estimate are yet to make up their minds. The latest YouGov poll put the undecided at just 4%, while an ICM telephone poll released the same day reckoned 17% of us have yet to choose. In assessing which of these figures is likely to be closer to the truth, it is worth considering the fact that YouGov’s surveys are sampled from their volunteer online panel. It is likely that such volunteer samples are more politically engaged than average and thus perhaps more likely to have decided how they will vote (analysis of SSA 2014 suggested that those who are less politically interested were particularly likely to be undecided).
But whatever the true number of ‘undecided’ voters, with the polls as tight as they are at present, their votes clearly matter a great deal to the outcome. So what are the issues that may sway them?
Since we first asked the now famous (or infamous, depending on your take) £500 questions, which asked people whether they would be for or against independence if they were £500 better or worse off, SSA has shown that economic issues are at the heart of the questions the Scottish public are asking themselves about independence. This remains the case – those who believe an independent Scotland’s economy will be better overwhelmingly support a Yes vote, while the vast majority of those who think it would be worse will vote No.
And for undecided voters, it is clear that one of the questions they remain undecided about is which of these two scenarios is the more likely – just under half (46%) of those interviewed for SSA 2014 who said they did not know how they would vote (even when pushed) said they did not know if Scotland’s economy would be better or worse under independence.
But with the polls so close, the campaigns cannot afford to ignore other issues that might influence undecided voters either – feelings of identity, as discussed above, but also questions of democracy and power. Several recent polls (from YouGov, ICM and Opinium) have highlighted the extent to which negative feelings about Westminster are a key reason people say they will vote Yes. And, according to Opinium at least, those who are undecided cite worries about Westminster control as among the key concerns they would have about staying in the Union.
So as campaigners hone the final messages they want voters to take to the ballot box, they will need to think very carefully about the balance they strike between issues of economics and those of power, and between messages of the head and those of the heart.