Although both of the polls published on Sunday, one from YouGov and one from Panelbase, suggested that the referendum race was too close to call, they painted a very different picture when it came to how far things had changed. Whereas YouGov were reporting a third poll in a row to register an increase in Yes support – cumulatively amounting to an increase of no less than 12 points in four weeks – Panelbase were saying that the position was much the same now as it had been as much as three months ago.
The latest poll today from TNS BMRB, another pollster that like YouGov has tended to report a relatively low Yes vote, is thus crucial is helping us to decide whether there really has been a substantial swing to Yes, or whether instead something unusual has happened to YouGov’s polls. Its answer is unambiguous; it also has detected a big swing to Yes.
The poll puts Yes on 38%, up six points on the company’s previous poll in late July and early August, while No are just marginally ahead on 39%. (Don’t Knows, always relatively high in TNS BMRB’s polls, are unchanged at 23%.) But that No lead is so small that once the Don’t Knows are eliminated, Yes and No are both on 50% (as is also the case of we only look at those who say they are ‘certain or very likely’ to vote). This represents an eight point swing to Yes, and is by far and away the highest Yes vote recorded by TNS in any of their polls.
As a result of today’s poll, our poll of polls, although still dated 5 September (as interviewing for TNS BMRB’s poll concluded on 4 September before that for YouGov’s), has been updated. On that measure Yes support has edged up again and now stands at 48%, with No on 52%. The referendum race is now clearly too close to call.
Today’s poll also shares some similarities with YouGov’s in terms of which groups of voters appear to have shifted most and least over the last month. Support for Yes has increased rather more amongst women and less well off voters and barely at all amongst older ones. The swing amongst women (once Don’t Knows are excluded is 11 points while amongst men it is only five. Equally Yes have gained 12 points amongst less well-off C2DE voters, but only four points amongst middle-class ABC1 ones. Meanwhile the swing amongst the over 55s is just two points.
However, one aspect of YouGov’s poll that is not replicated by TNS BMRB is the finding that there has been a particularly large swing amongst those who voted Labour in 2011. At 26% the figure was already rather higher than YouGov were reporting in early August (18%) and it is now, at 29%, just three points higher. Instead the most impressive gain has been amongst those who said they did not vote in 2011 (who of course constitute no less than half of all voters), amongst whom Yes support has increased (amongst those who have decided) from just over a third (35%) to just under a half (49%). The No side would, it seem, be unwise to assume that all it has to do to recover its position amongst ‘traditional’ Labour voters; rather it may need to recapture the much larger group of voters who have no particular party loyalty at all.
Although today’s poll shows no drop in the proportion of voters who say they do not know (even though the proportion who say they are certain to vote has leapt from 71% to 84%), overall it seems that voters now feel rather better about how much information they have to make a decision. When in June voters were asked to give themselves a mark out of ten to say how well informed they were, only just over a third (36%) gave themselves a mark of seven or more. Now well over half (58%) do so. Yes voters (76%) are particularly likely to feel they have enough information. Perhaps this is an indication that voters have been paying more attention in recent weeks, and what they have seen and heard during that time has persuaded some of them that maybe independence is not such a bad idea after all?