Who wants to take on the “Scottish” Daily Express?

As I’ve pointed out before, it’s sensible for bloggers to limit the number of complaints they personally lodge with the press regulator. However, the “Scottish” Daily Express website has just published another wildly misleading article about Scottish political polling – this time about the seats projection that some random dude on Twitter stupidly applied to a tiny, incorrectly-weighted Scottish subsample from a GB-wide Redfield & Wilton poll.

Now, we know from past experience that the press regulator (a non-independent body essentially run by the press themselves) sets an extremely high threshold to even consider complaints about misleading claims relating to polls. If there’s any convoluted excuse to be made about something being a matter of interpretation, that’s deemed to give the publication a free pass to mislead to its heart’s content. So to have a realistic chance of getting a complaint to stick, you need to find a flat-out untruth.

In spite of the attempts of the Express to cover themselves with caveats about polling methodology, it seems to me there is one particular sentence in the article that may be directly inaccurate. It’s essentially a sub-headline, and states:

“The Scottish section of the poll – although it is based on a weighted sample of 180 people – provides yet more evidence that the Nats are falling behind a resurgent Labour party”

The subsample actually shows the SNP on 34%, three points ahead of Labour, who are on 31%. I’m struggling to see how that shows evidence that “the Nats are falling behind” Labour. Presumably the Express would fall back on the incredibly thin excuse of an idiotic seats projection made by a third party, but then they would have to explain the phrase “yet more evidence”. Where is this other considerable evidence that Labour are actually ahead of the SNP? There’s in fact vast evidence from all recent full-scale Scottish polls that the SNP remain well ahead of Labour, both in terms of vote share and projected seats.

Additionally (although I’d put less stress on this point), the “weighted sample of 180 people” did not involve 180 “people” as such. Only 143 actual people were interviewed for the Scottish subsample, with the results adjusted to count them as if there were 180. By the time you get to the section of the sample that was used to produce the results reported in the article (ie. once Don’t Knows and Won’t Votes were stripped out), there were just 121 real people, upweighted to count as 137.

I’m reluctant to lodge a complaint myself, but I do very much feel a complaint is warranted. So if you’d like to complain to the press regulator about the inaccuracies in the IPSO article, you can find the necessary form HERE.