At first glance, the poll from TNS BMRB published tonight (conducted on behalf of a new organisation being founded by Sir Tom Hunter, with the aim of improving the quality of the data and the evidence in the referendum debate) presents a very different picture of the state of the referendum race than last week’s poll from ICM poll. Whereas last week’s poll put the Yes vote on 37%, just seven points behind No, this poll (which was actually conducted slightly earlier than the ICM exercise) puts the Yes vote at just 29%.

This, however, would be to miss the significance of this latest poll.

First, as we have repeatedly pointed out, TNS BMRB regularly find a much higher proportion of Don’t Knows than do other pollsters. This is primarily because the company asks people how they intend to vote when the referendum is held in September rather than what they would do if the ballot were to be held now. Mind you, as one might expect, that proportion has now dropped a bit, from 33% last month to 29% now – but that latter figure is still much higher than the 19% recorded by ICM.

To make sense of polls with such divergent numbers of undecided voters – and to get a clearer picture of the referendum result to which any individual poll is pointing – we should exclude the Don’t Knows from our calculations and identify what proportion of those who do have a stated view are currently inclined to vote Yes or No. On that basis the latest TNS BMRB poll translates into a 41% Yes vote.

Now, that is still markedly lower than the equivalent figure for last week’s ICM poll – 46%.  But then we should remember that TNS BMRB’s estimate was lower than ICM’s when the latter company had previously polled, in September last year. At the stage ICM were at 40%, TNS BMRB, on 36%. In a referendum campaign in which different polling companies have repeatedly been producing rather different estimates of the relative strength of the two camps, TNS BMRB’s estimates have usually been towards the pessimistic end of the spectrum so far as the Yes side is concerned.

Thus, the result of this poll represents a 5 point swing to Yes since last September, almost identical to the 6 point swing that ICM registered (once the undecideds were excluded). Moreover, like the ICM poll, this poll has uncovered a record high level of support for Yes too – it is the first time that TNS BMRB have put the Yes vote above the 40% mark.

Of course not even two swallows necessarily make a summer. However, all in all, as many as six polls have now been conducted since the Scottish Government’s White Paper was published. All six have reported at least some increase in Yes support – on average one of three points. Such a swing is far from dramatic and still leaves the Yes side with an average level of support (in this particular set of polls) of just 39%. Even so, it represents the first discernible movement of opinion of any note in a referendum campaign that had hitherto looked remarkably stable.

Today’s poll also contains a question on people’s preference between independence, ‘devolution max’, and the status quo. Although a newly worded question, the findings are not wildly at odds with those of previous such exercises, including ones by TNS BMRB themselves. Devolution max emerges as narrowly the most popular option with 35% in favour, with 31% backing the status quo and 24% independence. Here it seems is another reminder, echoing that provided by the recent Scottish Social Attitudes survey, that the first preference of around one in three voters does not appear on the ballot paper.

At the moment relatively few of these ‘devo max’ voters are inclined to vote Yes. For every one that says they will do so, over three state they will vote No. But no less than 37% say they do not know how they will vote, and as a result they account for as many as 44% of all those who say they do not know what they will do in September. The absence of ‘devo max’ on the ballot paper evidently means they are finding it more difficult to make up their minds.  Now that the No side’s lead has narrowed, it may feel under somewhat greater pressure to reach out to these voters by stating more clearly than it has done so far, what in the way of more devolution might be on offer.

About the author

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