The excitement generated by the narrowing of the No lead in the polls has unsurprisingly persuaded new media organisations to enter the polling fray. On Friday The Guardian commissioned its first poll of voting intentions in the referendum. Today its stablemate, The Observer, has done the same. That paper regularly publishes Westminster voting intention polls conducted by Opinium Research, and thus it is to that company to which it has turned for its poll, even though the company has not previously undertaken any polling of referendum voting intentions. Consequently we have a late new entrant to the list of companies that are attempting to measure Scotland’s apparently febrile political mood.
According to the poll, amongst those who say they will definitely vote (no less than 90% of the sample), 45% say they will vote Yes, 49% No, while just 6% are undecided. Once the Don’t Knows are excluded this equates to Yes 47%, No 53%. So while this poll does not give us any direct evidence on how much the balance of opinion has changed recently, it would appear to affirm the message of most other polls that the referendum race is now very tight, albeit with the No side narrowly ahead.
The weighting strategy employed by Opinium is worth a brief note. The company does not weight its regular polls of Westminster voting intentions by how people said they voted in the 2010 election. In this poll, however, they have weighted both by how people said that they voted in that election and how they said they voted in the 2011 Scottish Parliament election. In so doing the company uncovers a now familiar pattern, which is that far too many people say they voted SNP in 2010 (as compared with the actual result) while there is something of a deficit in the proportion who say they did so in 2011. Both ‘errors’ have been corrected by Opinium’s weighting scheme, which implies that those who say they voted Labour in 2010 but SNP in 2011 will have been heavily upweighted. It remains to be seen whether this proves to be a better strategy than simply weighting by 2011 Scottish Parliament vote, as has become the norm amongst those other companies that weight their samples by how people say they voted in the past.
What no attempt has been made to do – and in this Opinium are from alone – is to weight the sample so that the reported level of abstention in 2010 and/or 2011 is closer to that of the official result. In fact the net effect of the weighting scheme is to reduce that proportion. As the company itself acknowledges this means that regular voters are overrepresented in their sample – and as we have noted previously there is a tendency for those who say they did not vote in 2011 to be more likely to support No (although whether that is true of this sample the published tables do not enable us to tell). In the meantime, we can but note that the net effect of Opinium’s weighting scheme is to increase its estimate of the Yes vote slightly, an outcome that is very much in line with that of other pollsters.
Although today’s poll does not give us any direct evidence on whether there has been a swing to Yes in recent weeks – a subject on which the regular polls disagree – it does contain some interesting evidence on people’s own self-reports as to whether or not they have changed their minds in the last three months. Those reports are consistent with the proposition that there has been a swing. As many as 27% say they have become more likely to vote Yes, while only 9% say they have become more likely to vote No. Just 42% of Yes supporters say they were always going to vote Yes, whereas 83% of No supporters claim to have been constant in their view. Doubtless some Yes supporters who say they are now more likely to vote that way are simply telling us that an existing conviction has got yet stronger, but one doubts that that can account for all the discrepancy in the apparent ability of the two sides to attract new voters.
Indeed, much of today’s poll paints a picture of a No campaign that is still ahead despite itself. Respondents were much more likely to nominate the Better Together campaign than Yes Scotland as the one that has been too negative (by 53% to 18%), disorganised (49% to 19% and lacklustre (54% to 14%). (The Yes campaign is, in contrast, more likely to be regarded as ‘unrealistc’ and ‘aggressive’.) Meanwhile, as YouGov identified last week, the No campaign no longer has any edge over the Yes side when it comes to credibility; just as many (44%) said that Better Together were untruthful as reckoned Yes Scotland were. Of course such perceptions are heavily influenced by people’s voting preference, but nevertheless, these figures indicate that many a No supporter has not been too impressed with what they have seen. For example, even No supporters are more likely to think the Better Together campaign (45%) has been lacklustre than they are its Yes Scotland counterpart (24%).
This picture extends to the personalities of the campaign. As many as 51% think Alex Salmond has done well in the campaign, while 48% say the same of Nicola Sturgeon. In contrast the best rating that anyone on the No side can manage is the 25% secured by Alistair Darling. Although the Yes side remain the underdogs in the referendum race, it certainly appears that its campaigning efforts have secured widespread respect. Unionists might well be advised to reflect that mood should they still manage to come out top on Thursday.