Thursday, February 20, 2014

Pro-independence campaign make huge breakthrough in new Survation poll

As far as I can gather, there may be as many as three new referendum polls out tonight, although we only have figures for two so far.  Let's start with Survation's second contribution to this campaign, which shows the No camp's lead being cut by more than half -

Should Scotland be an independent country?

Yes 38% (+6)
No 47% (-5)

The fieldwork is bang up to date (the first two days of this week), so the figures will take full account of any backlash against George Osborne's ill-judged intervention on the currency issue.  However, before we all get too carried away, it should be remembered that we can't make a truly like-for-like comparison with the last Survation poll, because John Curtice was so scathing of the methodology that Damian Lyons-Lowe (to his great credit) more or less immediately committed himself to making a significant change, which we can assume has now been implemented. Even allowing for that, though, it still seems likely that there has been substantial movement to Yes.

UPDATE (11.40am) : Survation have confirmed the change in methodology, and that they are now weighting by recalled Holyrood vote from 2011 rather than recalled Westminster vote from 2010.  However, even on a direct comparison between the raw unweighted figures in the last poll and in this one, the No lead has slumped by some 4.3%, suggesting that there has been a significant backlash against the Osborne speech.  The methodology change should of course be warmly welcomed, as it means that the Yes vote will no longer be understated by a misconceived weighting procedure.  It's interesting that Survation make the point that YouGov have still not come wholly into line with this best practice.

This is how the picture looks with Don't Knows excluded -

Yes 45% (+7)
No 55% (-7)

Next up is TNS-BMRB's second published poll of the year (although mysteriously they've added a previously unheard-of one from early January to their trends graphic), which is somewhat less significant because the fieldwork was conducted well before Osborne's foray north. The headline figures show a static picture since the most recent poll, indicating that the Yes campaign have retained all of the extraordinary gains they've made over the last few months (the No lead has gradually fallen from 22 points to 13 since the late summer/early autumn) -

Should Scotland be an independent country?

Yes 29% (n/c)
No 42% (n/c)

With undecideds excluded, it works out as -

Yes 41% (n/c)
No 59% (n/c)

The real excitement, though, lies in the figures that only include respondents who are certain to vote. On this measure, the No lead has now slumped to single figures for the first time since TNS-BMRB introduced their monthly survey -

Yes 35% (+3)
No 44% (-5)

With undecideds excluded, it's -

Yes 44% (+4)
No 56% (-4)

And just to reiterate - this is progress that was made before Osborne even opened his mouth!

I started this post by saying I thought there may be three new polls out tonight - that was largely based on a tweet from a BBC journalist who claimed that there was a poll in the Daily Mail showing the No lead dropping from 12 to 9. As I write this I'm beginning to think he was probably getting mixed up, because the only pollster that had previously shown the No lead on 12 was Panelbase, and of course it's the Sunday Times that use Panelbase, not the Daily Mail. So it looks like we may have only two polls tonight, but I think that will do us quite nicely to be getting on with!

* * *


As you'd expect, the pro-independence campaign's spectacular advance in the Survation poll has helped Yes reach its best ever position in the Poll of Polls. This is the tenth time the figures have been updated, and the seventh update to show a swing towards Yes (two showed a swing to No, and one an unchanged position).

MEAN AVERAGE (not excluding Don't Knows) :

Yes 34.7% (+0.8)
No 48.3% (-0.7)

MEAN AVERAGE (excluding Don't Knows) :

Yes 41.8% (+0.9)
No 58.2% (-0.9)

MEDIAN AVERAGE (excluding Don't Knows) :

Yes 42.0% (+1.2)
No 58.0% (-1.2)

(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 - Angus Reid, YouGov, TNS-BMRB, 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.)

For the very first time since the Poll of Polls started, the median average now shows a marginally smaller gap than the mean average does. This is caused by Survation rocketing up the rankings from having previously been the second-least favourable pollster for Yes to now being the second-most favourable. For the time being, Panelbase have slipped to being only the third-most favourable for Yes - so much for it being "Panelbase versus the field", as Alex Massie used to say!

On the headline figures, the No lead has fallen to just 13.6%, meaning that Yes require a swing of 6.8% to draw level.


  1. James, I've had a little look around and the TNS poll from early January was the one commissioned by the BBC (which the full results tables for have now been released)

  2. James, Who commissioned the Survation poll?

    They show a gap of 9% between Yes and No, so maybe this is what the BBC man was referring to?

    We know how diligent the Beeb have been to make sure they are reporting accurately, when referring to the referendum, so... oh wait a minute!!!

    Maybe it's just more crap from the BBC!

  3. The game is afoot methinks
    take Scottish Skiers assertion and the distance is not even 6.8% I would suggest we are reaching parity if not in front, all you have to do is look at the fear in their eyes to see its closer than they are letting on.

  4. James re the Survation poll;

    Data were weighted by gender, age, region and 2010 general election vote.

    So it looks like we can compare with their previous poll.

  5. Marcia : When I checked 10 mimutes ago, Survation still hadn't released any details. The datasets that have been posted on Twitter and Wings are from the last Survation poll earlier in the month.

  6. James the person who post a link on WoS to the data table got it wrong it. According to John Curtice it has been weighted to 2011 so he will have seen the tables.

  7. Here are the Survation tables:

  8. Patrick : Sorry, I overlooked your question. The Survation poll was commissioned by the Mail, so that may well have been what the BBC man was talking about.