Wednesday, April 16, 2014

Dramatic TNS-BMRB poll shows pro-independence campaign closing the gap to the lowest level so far

This is hugely significant, more so than any Panelbase or even ICM poll. TNS-BMRB are different because they are one of only two non-online pollsters that have been active in the referendum campaign so far. The other one of the two, Ipsos-Mori, has persistently shown a much higher No lead than any of the online pollsters, so if TNS-BMRB were corroborating those results it would be a matter of some concern for the Yes campaign. But thankfully that isn't happening. The latest TNS-BMRB poll has just been released, and it's very much in line with the trend shown by the online pollsters.

Should Scotland be an independent country?

Yes 29% (+1)
No 41% (-1)

The changes from last month may look modest, but as in the last YouGov poll they're enough to reduce the No lead to its lowest level of the campaign so far. The last TNS poll to show a No lead of lower than 12% was way back in early 2012 before the campaign got underway, when a radically different question and methodology was being used. The No lead is now less than half of what it was at its peak.

Because TNS always report a much higher number of undecided voters than other pollsters, the best way of making sense of the figures is to strip the Don't Knows out, which produces a position of -

Yes 41% (+1)
No 59% (-1)

That's more or less in line with the record high of 42% that YouGov found in their most recent poll, although of course it's still a few points short of the 44-47% Yes vote currently shown by Survation, ICM and Panelbase.

The figures for respondents who say they are certain to vote (which some pollsters would use as the headline numbers) also show a slightly narrowing lead -

Yes 33% (+1)
No 46% (n/c)

That isn't actually the lowest the No lead has been on that measure - it fell to as low as 9% in a poll conducted in late January, and had already been at 10% a few weeks earlier. However these figures have proved to be much more volatile than the headline numbers, probably due to the low sample size once you strip out the huge number of undecideds and people who are less than 100% certain to vote. So I doubt if there's any great significance to the fact that the gap isn't quite the narrowest it has been so far. The salient point is that the No lead among certain voters has slipped by 9% since September, and by 5% since October.

With Don't Knows excluded, the position among certain voters works out as -

Yes 42% (+1)
No 58% (-1)

One of the irritating things about TNS-BMRB is the long gap between the fieldwork taking place and the results being reported. Incredibly, this poll was conducted between the 21st of March and 2nd of April, which means that it entirely predates the most recent Panelbase and Survation polls. So in a sense we're just looking into history with these numbers - they don't get us any closer to knowing whether the pro-independence campaign have been making further gains in the most recent period, or simply consolidating the gains they've already made. However, this poll certainly bolsters the impression that, at the very least, Yes gained a little more ground over the course of February and March.

* * *


Tonight's update of the Poll of Polls equals the record for the lowest No lead to date, and indeed it's only the rounding that cheats us of a new record being set.

MEAN AVERAGE (not excluding Don't Knows) :

Yes 35.4% (+0.1)
No 47.9% (-0.1)

MEAN AVERAGE (excluding Don't Knows) :

Yes 42.5% (+0.1)
No 57.5% (-0.1)

MEDIAN AVERAGE (excluding Don't Knows) :

Yes 42.0% (n/c)
No 58.0% (n/c)

(The Poll of Polls is based on a rolling average of the most recent poll from each of the pollsters that have been active in the referendum campaign, and that adhere to British Polling Council rules. At present, there are seven - YouGov, TNS-BMRB, Angus Reid, Survation, Panelbase, Ipsos-Mori and ICM. Whenever a new poll is published, it replaces the last poll from the same company in the sample. Changes in the Poll of Polls are generally glacial in nature due to the fact that only a small portion of the sample is updated each time.)

You may have seen that the Financial Times have introduced their own Poll of Polls for the referendum - the first chart can be found HERE. I know the trend they're showing looks thrillingly good, but unfortunately it's slightly misleading. Unlike the method I use on this blog, the FT are using a rolling average of the last seven polls to be published, regardless of which pollster conducted them. That means the current sample of seven polls contains five from the more Yes-friendly companies, including three from Panelbase alone. If the next seven polls come mainly from No-friendly pollsters, the FT are bound to report a substantial increase in the No lead even if the individual polls all show the complete opposite to be happening!

So I do think my own method is more appropriate for this very unusual campaign where the results from different polling companies are diverging to a highly significant degree. It does have one big disadvantage, though, which is that I don't know what to do about Angus Reid - I'm still including their most recent poll in the sample, even though it was conducted last summer. If they don't make a reappearance soon I'll have to remove them, but I'm reluctant to do it because it will make the long-term trend figures less meaningful. Talking of those figures, here is the latest update -

The No campaign's lead in the Poll of Polls headline figures :

Sep 2013 - 20.2%
Sep 2013 - 20.0%
Sep 2013 - 18.4%
Oct 2013 - 17.9%
Oct 2013 - 17.5%
Oct 2013 - 17.4%
Nov 2013 - 17.5%
Dec 2013 - 17.1%
Dec 2013 - 16.3%
Dec 2013 - 16.2%
Dec 2013 - 15.8%
Jan 2014 - 14.2%
Jan 2014 - 14.8%
Feb 2014 - 14.8%
Feb 2014 - 14.7%
Feb 2014 - 15.1%
Feb 2014 - 13.6%
Feb 2014 - 14.0%
Mar 2014 - 14.0%
Mar 2014 - 14.3%
Mar 2014 - 14.3%
Mar 2014 - 13.6%
Mar 2014 - 12.9%
Mar 2014 - 13.0%
Mar 2014 - 12.5%
Apr 2014 - 12.6%
Apr 2014 - 12.7%
Apr 2014 - 12.5%


  1. I keep wondering if there is a tactic for holding back details of polls? This is the second poll that has had later polls published so make them out of date. However it is good to see the trends in the polls from the different firms.

  2. I did a Yougov poll about a few days ago wonder when it will publish

  3. Marcia : TNS are fairly consistent with their delays, so it's probably because they're an old-fashioned face-to-face pollster (although Anthony Wells suggested that it still ought to be possible for them to do it quicker).

    Rod : Should be quite soon, unless it was an internal or testing poll (in which case it might not be published at all).

  4. Just 30% 'definitely decided' on No.

    Sounds about right.

  5. Meanwhile on a website far far away (you know the one) it's going to be 80% No and Moira Salmond is fair game for any sleazebag to abuse. Well done the fair and unbiased moderators.

  6. LAST 7 usually the best way to do it because that usually means a few months polling.

  7. Not really - the last seven polls only cover a period of just over a month. In any case, the real problem is not timescale, but the randomness of whether the average goes up or down depending on whether there are a particularly high number of polls being published by Yes-friendly or No-friendly pollsters.

  8. Looking through the tables, the general observations I have are:

    -TNS are showing the lowest NO lead so far, with the lead dropping 7% in 6 months.

    -YES has a 4% lead in the DE category and is very close among under 55s

    -21% have definitely decided to vote YES, 30% definitely NO, making this race even more open that we thought. Gives us reasons to be optimistic that there could be a turnaround in opinion similar to what we see after all those university debates

    -Only 44% of those who intend to vote for independence voted SNP in 2011