John Ward – Scottish Indyref : ICM Puts Yes Vote Ahead By Seven Points WITHOUT Removing The Don’t Knows – 14 September 2014

JohnWNeck and neck, but signs that the undecided 10% are veering towards Yes

There’s another new internet-based poll on the Scottish Independence vote out today from ICM. It shows the Yes campaign on 49%, No on 42%, with 9% of don’t knows. I hate to be an anorak about this, but as consumer research used to be my metier, I feel it important to make telling comment.

Previously, there have been two critiques of such internet polling: (1) it isn’t representative; and (2) the Barker Poll of last week left out the Don’t Knows.This latest ICM study casts doubt on both those viewpoints.

As for internet polling being unrepresentative, you would be hard-pushed to find anyone under 70 these days who has never used the Web: in fact, 95% of all people aged over 65 in the UK do so. So as I said in the last Slogpost on this subject, to call such a survey unrepresentative is daft. On the other side of the coin, if pc users are near-universal, is the No campaign suggesting that it’s ‘lead’ is entirely down to Neolithics who aren’t computer literate? In which case, boy do we need to worry about those voting No in such a manner as to be decisive: some of them may well all be prototypes for Rab C Nesbitt that never quite made it off the drawing board.

More seriously, the ICM poll is among internet users, and its sample size is relatively low: at 705 respondents, the margin of error is + or – 3.8%. So the Yes/No score could be 46/45 – too close to call – or it could of course be 53/38. To be frank, as I posted last week, I think we’re looking at ‘too close to call’….with around 1 in 10 voters still not sure.

Turning now to the second bitch with the Barker poll, this new ICM survey includes Don’t Knows. Put another way, not only is this race too close to call – the Don’t Knows could easily turn it decisively one way or the other. Furthermore, the idea that ‘forced choice’ made the Barker poll worthless is given the drubbing it deserves: here is another methodology saying the same thing: neck and neck.

Meanwhile, a Survation Poll shows the No vote clearly in the lead, but with the Undecideds still at 10% and thus sufficient to wipe it out.

The ICM poll was commissioned by the Sunday Telegraph. So as it’s their poll, you’d think they’d lead with that one – not the Survation one.

Is the Telegraph leading with it today? It is not. Instead both polls are quoted – and given less space – although to be fair to the Biffo Barclay Boys, they are at least taking The Slog’s lead and using the phrase ‘too close to call’. (And the predictable demise of David Haines was more than enough to hog the front page).

I would hazard a guess that the Torygraph commissioned an internet survey including don’t knows….in the hope that it would give a different result to the Barker poll. But ‘net’, it gave exactly the same result.

What’s more, if you place all polling done so far in series, the No Vote is not winning the battle of the Don’t Knows – it’s losing it: the margin is narrowing.

There is an old adage which asserts, “There are lies, damned lies, and statistics”….but I have long maintained that this is quite untrue. There are only statistics, and discernible mendacity in interpreting those statistics.

In this case, we have one ‘result’ confounding the bollocks about leaving out Don’t Knows…and then more ‘results’ showing that, as we go down to the wire, signals that the Yes bandwagon is enthusing more of the neutrals.

Let’s face it, the mendacious interpretation of statistics has another much shorter name: propaganda.

Yesterday at The Slog: It’s dog eat dog in Scotland / link to original article


Comments are closed.