More Polls: Still Close!

Yesterday evening saw the publication of what are almost the final polls of the campaign – there is one more from Ipsos MORI due to appear in today’s Evening Standard. Between them they affirmed the impression created by the polls released on Monday evening that the outcome of the referendum is on a knife-edge but with the No side marginally the favourites.

Perhaps the most important of the four polls in question came from Ipsos MORI and was conducted for STV. Throughout the campaign the company has been one of those that has painted a relatively pessimistic picture of the Yes side’s prospects. In their last poll, published on the evening of the first leaders’ debate at the beginning of August. It put Yes on just 42% once undecided voters were excluded.

However, the impression created by the latest poll is very different. Yes are estimated to be on 47%, No on 49%, with just 5% stating that they are undecided. Once the small number of Don’t Knows are excluded that equates to a Yes vote of 49% and a No one of 51%, representing a seven point swing to Yes since the company’s previous poll. The poll thus corroborates the large swing to Yes since the first half of August that has previously been identified by both YouGov and TNS BMRB and which has fundamentally changed the mood of the campaign and the apparent prospects of the two sides.
We should though note one or small differences between the basis on which Ipsos MORI have calculated their headline figures in this poll and the approach which it has adopted in its previous polls. First the company has asked its respondents to confirm that they are registered to vote, though unsurprisingly as many as 98% said that they were. Second, although the figures are based only on those who say they are certain to vote (who, however, constitute no less than 95% of this sample) they included the responses of those who initially said they did not know how they would vote but then when asked which way they were more inclined responded either Yes or No. This procedure inevitably reduced the proportion of Don’t Knows that were reported, but as it happens made no net difference to the relative balance of Yes and No support.

The second poll came from YouGov and was conducted for The Times and The Sun. It put Yes on 45% and No on 49%, with just 6% saying they don’t know or are undecided. Support for No is down a point as compared with the company’s previous poll conducted a week ago. Once the Don’t Knows are excluded Yes are on 48% (and No on 52%), unchanged on the company’s previous reading. The company has not made any adjustment for reported propensity to vote, but as it happens the relative balance of Yes and No support is the same amongst the 90% who said they were certain to vote as it was in the sample as a whole.

The third poll came from Panelbase, the company that for most of the campaign consistently produced the highest estimate of the Yes vote. It put Yes on 45%, No one point drop in Yes support and a three point increase in the No tally as compared with the company’s previous poll conducted last week. However once the Don’t Knows are excluded the Yes vote at 48%, (with No on 52%) is only one point down on that previous poll. Even so, the Yes side must be disappointed that their favourite pollster has in the event failed to put it at least narrowly ahead.

The final poll was undertaken by Survation for the Daily Record. For this exercise the company reverted once more to polling by phone rather than over the internet, as they had done previously in a poll conducted last week for the Better Together campaign. It put Yes on 43% and No on 49% while 9% indicated they were still undecided. The Yes vote is up one point while the No vote is the same as compared with the company’s previous phone poll. Once the Don’t Knows are excluded Yes are on 47% (and No on 53%), up a point on the previous phone poll.  Given that the Yes tally in the company’s final internet poll was 48% and was repeatedly at 47% in the company’s polls during the campaign, the exercise appears to have confirmed the impression created by previous switches to phone polling by both Survation and ICM that the two forms of polling produce much the same estimate of Yes and No support.

Thus in the last 24 hours we have had seven opinion polls, all of which have put Yes narrowly behind on between 47%-49% – a remarkable convergence between the pollsters given the persistent divergence between them during early phases of the campaign. Individually the polls are saying that, bearing in the mind the possibility that their estimate could be somewhat inaccurate as a result of the chance error to which all such exercises are subject, we cannot rule out the possibility that Yes could win 50% + 1. However, it should be appreciated that when seven polls all point to more or less exactly the same outcome, the chances that Yes will win more than 50% is much lower – assuming, that is, that the polls are not all systematically in error. Meanwhile it will come as little surprise that our poll of polls now stands at Yes 48%, No 52%, a one point drop in Yes support as compared with our previous calculation.

If the apparent swing to Yes in much of August and early September has indeed stopped somewhat short of what it will require for victory tonight, it is clear that there are two key demographic groups that will have denied it victory. The first is women, only 43% of whom on average said they were going to vote Yes in the final regular poll of each company – well down on the 52% of men who sais they intended to vote Yes. The second is older people; all companies put the level of support for Yes at below 40% in the oldest age group that they identify in their tabulations.

Meanwhile so far as the issues are concerned, if the Yes side does lose it will probably have done so not least because it never managed to persuade a majority of Scots that the country would be more prosperous under independence. YouGov find in their latest poll that only 35% think Scotland will be economically better off under independence while as any as 47% reckon it would be worse off. Although smaller than anything registered by YouGov before September, that net pessimism of 12 points compares with an equivalent figure of just two in the now famous YouGov poll ten days ago than put the Yes side narrowly ahead. It may well be that the many warnings of the economic consequences of independence that came from No campaigners and many a financial institution in the last ten days of the campaign has had its desired effect.

This impression that the Yes side have failed to do enough to win the argument about the economic case for independence is underpinned by a finding from Panelbase. Those who have followed the referendum campaign from its beginning will be aware of a famous finding from Scottish Social Attitudes that a majority would support independence if they thought they would be £500 a year better off as a result.  However, Panelbase tell us that just 13%, have been convinced that that is what would happen, while as many as 33% believe they would be £500 a year worse off.  If Scotland does vote today to stay in the Union today, it appears it will have done so because that is where a majority have concluded that their nation’s and their own material interests lie.

Topics: The Scottish independence referendum

About the author

John Curtice is Professor of Politics at Strathclyde University, Senior Research Fellow at ScotCen, and Chief Commentator on the What Scotland Thinks website.