Sunday, June 15, 2014

Pro-independence campaign close the gap to just 7% in dramatic poll from "gold standard" ICM

The second referendum poll of this evening is now out, from the firm that is often regarded as the UK's "gold standard" polling organisation.  It shows the No lead almost halving from 12% to 7% -

Should Scotland be an independent country?

Yes 36% (+2)
No 43% (-3)

With Don't Knows excluded, it works out as -

Yes 45% (+3)
No 55% (-3)

When the datasets are released, the first thing I'll be looking out for is whether Yes were unlucky not to be rounded up to 46%, which of course would also have meant No being rounded down to 54%.  A rough calculation suggests that may well have been case.  If so, it would be a repeat of the March ICM poll, when Yes missed out on being rounded up to 46% by the tiniest fraction imaginable.

So we've had three polls since Wednesday evening, and every single one has shown Yes closing the gap by a significant degree.  It must be conceded now that, by any rational measure, the chances of Scotland becoming an independent country have increased markedly of late.  However, it's true that there is a difference between Panelbase and Survation on the one hand, who are both showing the No lead at its lowest point ever, and ICM on the other hand who are showing a result that is within their previous normal range, even though the No lead has slumped since the most recent poll.  I'm sure John Curtice will be seizing upon that fact to make the case that we do not yet have absolutely conclusive proof that there has been a renewed momentum towards Yes in recent weeks.  But the problem with that argument is that ICM keep changing their methodology, and it's therefore a touch meaningless to talk about a normal range - if they had been using their current methodology all along, it's absolutely impossible to know what they would have been showing in February or March.  All we can really say for sure is that they're back to showing an extremely close race, and in that sense they're providing corroboration for Panelbase and Survation.

We can also officially welcome ICM back into the fold of relatively Yes-friendly pollsters.  As Calum Findlay pointed out a month ago, in their last poll ICM somehow went at a stroke from being the most Yes-friendly pollster of the lot to being one of the most No-friendly.  Indeed, if they hadn't introduced turnout weighting in that poll, they would have showed a Yes vote of just 40% after rounding - exactly the same as Ipsos-Mori reported a couple of weeks ago.  That poll just smelt all wrong from start to finish, and although it's possible that it showed an inflated No lead due to an extreme case of margin of error noise, my gut feeling is that it had more to do with the utter folly of experimenting with an introductory question that would have affected respondents' mindsets before they reached the referendum question.  Fortunately, we already know that the introductory question has been removed from tonight's poll, so in that sense there's no reason to question the results.  However, when the datasets arrive it'll certainly be worth checking whether ICM have once again significantly under-represented men and Scottish-born respondents, and over-represented English-born respondents.  If they have, that in itself might be sufficient to explain the remaining small difference between the 45% Yes vote they are reporting, and the 48% Yes vote reported by Panelbase.

Where are we left now in terms of the overall state of play?  As always (and this will presumably continue to be the case all the way up to polling day), there is huge uncertainty on what the true numbers on the ground are - the estimates of the Yes vote range from 40% to 48%.  It's not even the case that there is a broad consensus being contradicted by one or two outliers - the six polling firms are in fact spaced out reasonably evenly along that range.  Because so many polls are reliant on volunteer online panels, it's particularly important to pay attention to the two firms who actually go and seek people out in the 'real world' - and although it's true that both Ipsos-Mori and TNS-BMRB are broadly on the No-friendly end of the spectrum, there's actually an important contradiction between the two.  Ipsos-Mori report a Yes vote of 40% regardless of whether a turnout filter is applied, but by complete contrast the turnout filter makes a huge difference to the TNS results.  Among those certain to vote, Yes is at 44%, and if you add in 'undecided leaners' the figure is 45% - exactly the same as ICM are reporting tonight.

And as we all know in life, where you are right now is a hell of a lot less important than where you're going.  Incredibly, it's now the case that five out of six BPC-affiliated pollsters have shown a decrease in the No lead in their most recent poll - the only exception is TNS-BMRB who showed a static position.  Four out of the six firms have shown the No lead at its outright lowest point of the campaign so far in their most recent poll, while a fifth (TNS-BMRB again) showed it at its joint lowest point.  Well, we know McDougall's minions can be very creative in serving the forces of darkness, but I think even they'd struggle to reconcile those trends with a repeat of the notorious "Momentum Is With No" graphic!

*  *  *

I've just caught up with John Curtice's analysis of tonight's polls, and although it's very fair and detailed, I find it most interesting for what it doesn't say than for what it does.  You can be 99.99% certain that Curtice will already have seen the full datasets for the Yes Scotland-commissioned Panelbase poll, which means that if there had been any funny business with leading wording or with the question sequence he would have mentioned it.  No objection is raised at all, meaning that we can regard the 48% figure as robust - Yes really have reached their highest level of support of the campaign so far with Panelbase.

Curtice also reveals that the headline Panelbase figures (ie. not excluding Don't Knows) are -

Yes 43% (+3)
No 46% (-1)

What's striking about that is just how high the raw Yes vote is - it's 2% higher than in any previous Panelbase poll that didn't use an unusual question sequence.

*  *  *


The last update of the Poll of Polls showed the No lead returning to a position very close to the previous all-time low, so it was always inevitable that two polls showing substantial swings to Yes tonight were going to see a new record being set by a good distance.  As it turns out, the swings are just about big enough to take the lead below the psychological 10% threshold on the average that takes account of Don't Knows.  As recently as September of last year, the No campaign had an advantage of 21.6% - more than half of that has now been wiped out.

MEAN AVERAGE (excluding Don't Knows) :

Yes 44.1% (+0.9)
No 55.9% (-0.9)

MEAN AVERAGE (not excluding Don't Knows) :

Yes 36.8% (+0.8)
No 46.7% (-0.6)

MEDIAN AVERAGE (excluding Don't Knows) :

Yes 43.8% (+1.5)
No 56.2% (-1.5)

(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 since September 2013, and that adhere to British Polling Council rules. At present, there are six - 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.)

In case you're wondering, ICM are solely responsible for the bigger shift in the median average, which is currently calculated as the mid-point between ICM and YouGov.

Here are the long-term trend figures, with the updates prior to Easter recalculated to exclude the inactive pollster Angus Reid...

The No campaign's lead in the Poll of Polls mean average (not excluding Don't Knows) :

Sep 2013 - 21.6%
Sep 2013 - 21.4%
Sep 2013 - 19.4%
Oct 2013 - 18.8%
Oct 2013 - 18.4%
Oct 2013 - 18.2%
Nov 2013 - 18.4%
Nov 2013 - 18.0%
Dec 2013 - 17.0%
Dec 2013 - 16.8%
Dec 2013 - 16.4%
Jan 2014 - 14.4%
Jan 2014 - 14.2%
Jan 2014 - 14.2%
Jan 2014 - 15.2%
Feb 2014 - 15.0%
Feb 2014 - 15.5%
Feb 2014 - 15.5%
Feb 2014 - 13.7%
Feb 2014 - 13.3%
Feb 2014 - 14.2%
Mar 2014 - 14.2%
Mar 2014 - 14.5%
Mar 2014 - 14.5%
Mar 2014 - 14.7%
Mar 2014 - 13.8%
Mar 2014 - 13.0%
Mar 2014 - 12.5%
Apr 2014 - 12.5%
Apr 2014 - 12.7%
Apr 2014 - 12.7%
Apr 2014 - 12.3%
Apr 2014 - 11.4%
May 2014 - 11.2%
May 2014 - 11.2%
May 2014 - 11.5%
May 2014 - 13.3%
Jun 2014 - 12.1%
Jun 2014 - 12.1%
Jun 2014 - 11.3%
Jun 2014 - 9.9%

It's been a while since I bothered to do an average for online pollsters only, but this somehow seems like quite a good moment to revive the tradition...


Yes 45.8%
No 54.2%

MEAN AVERAGE OF ONLINE POLLSTERS (not excluding Don't Knows) :

Yes 38.8%
No 46.0%


Yes 46.3%
No 53.7%

Don't forget the Generation Yes fundraiser, which now only has a few hours left to run.


  1. This comment has been removed by the author.

  2. Great news! We're edging our way to victory!

    And thanks for the plug, James!

    People can find out more about us at our website - sign up yer young 'uns!

    - GY

  3. James,

    I have posted this on WoS.

    Three times as many people trust Holyrood to make decisions in the interests of Scotland as opposed to politicians at Westminster, a poll has revealed.

    When asked who they trusted most to make decisions in Scotland's best interests, 52% of Scots said Holyrood, while 17.3% said Westminster, according to the study by Survation.

    When the 30.7% who said they did not know were excluded, 75% trusted Holyrood more while 25% preferred Westminster.

    A total of 1004 people were questioned for the research, which was commissioned by the Better Nation political blog in partnership with the Daily Record and Dundee University's 5 Million Questions referendum project.

    Amongst those who voted for the SNP in the 2011 Scottish Parliament election, 82.3% said they trusted Holyrood more to make decisions in the best interests of Scotland.

    More than half (51.9%) of Labour supporters also preferred the Edinburgh parliament over that in London.

    Almost three-fifths (57.9%) of those who backed the Tories in 2011 said they trusted Westminster more to make decisions in Scotland's best interests.

  4. Excellent post Marcia.

    As James already knows (he's probably bored of hearing it from me by now, LOL) if there is ONE thing I have stressed repeatedly and at length again and again, here and indeed on other sites for years, it is that the Independence Referendum will quite obviously boil down to TRUST.

    And when it comes to trust the westminster parties are well and truly f***ed.

  5. There may be a YouGov poll coming out soon (unless it already has & I missed it). As well as the indyref question asked for things like country of birth, and most intriguenly who you wanted to win the World Cup. Possibly that's the question to provide the headlines?

  6. Steve : YouGov seem to do a lot of internal polling for the No campaign, but I doubt if they would have bothered with the populist question about football unless this one was intended for publication. The fieldwork will be more up to date than ICM or Panelbase, so if it does see the light of day it'll be the first proper sign of whether the Lally/Rowling nonsense has had any impact.

  7. Apparently that panelbase, as per survation, points towards a majority for Yes at the prospect of more Tory rule. Interestingly, Ed Miliband in white knight garb on a devo nano horse doesn't do much better at 50Y/50N.

  8. James
    Could you explain the conversions from Y,N,U, to Y,N excluding U?
    Is it just a pro rata conversion of the Y,Ns, 50/50, or is it from data in the actual polls?

    And I have read elsehwere that Us r currently swinging 70% Y 30% N which if correct would change the Y/N excluding U quite dramatically

  9. No need for the "vote No to exterminate Scotland" campaign to panic.
    Christina D aka Fitalass over on Political Racism says that The SNP have failed to convince the young, old and women to vote Yes.

    I don't know why we're even bothering since it's such a forgone conclusion.

  10. James : The headline figures and the figures excluding Don't Knows are calculated in the same way - from the raw data. If you "converted" the headline numbers in the ICM poll, you'd actually end up with Yes 46%, No 54%.

    I'd take the 70/30 thing with a pinch of salt, but if it's true it refers to previous undecideds who have now been converted, not to current undecideds.

    Anon : It's only a few months since Fitalass was predicting a Yes vote of between 20% and 30% - I do hope she paid a visit to her local bookies!