Thursday, September 28, 2017

Davidson's despair deepens as latest batch of subsamples shows the SNP ahead across the board

I'm still in foreign climes, so I haven't been staying on top of the Scottish subsamples from Britain-wide polls.  But here's a round-up of the subsamples that have been published since my last update...

ICM: SNP 38%, Labour 27%, Conservatives 23%, Liberal Democrats 6%, Greens 5%, UKIP 2%

YouGov: SNP 36%, Labour 32%, Conservatives 22%, Liberal Democrats 8%, Greens 1%, UKIP 1%

Opinium: SNP 40%, Conservatives 31%, Labour 20%, Greens 6%, UKIP 2%, Liberal Democrats 1%

Ipsos-Mori: SNP 48%, Conservatives 23%, Labour 19%, Liberal Democrats 7%, Greens 1%

Survation: SNP 41%, Labour 23%, Conservatives 23%, Liberal Democrats 12%, UKIP 1%

There are two omissions from that list.  I can't find the datasets from the BMG poll, and there was another poll from Survation where there doesn't seem to be any Scottish figures in the datasets.  If anyone can fill in those gaps, please let me know.

As far as the five listed above are concerned, obviously there's no consistency on how far the SNP are ahead, or which party is in second place.  But it's really striking that the SNP are ahead right across the board, just as they were in the last batch of subsamples.  We seem to be getting back to the point where it will be quite unexpected to see a subsample with the SNP behind Labour or the Tories, whereas during the summer it was a fairly frequent sight.  This lends some support to my theory that the full-scale Scottish poll from Panelbase (showing an enormous SNP lead) took us by surprise not because the subsamples during the summer had misled us, but because the political weather has since changed, and the SNP have recovered from their post-election mini-wobble.  That reality now appears to be showing up in the subsamples, just as it did in that sensational full-scale poll.

The snag with my theory remains the same as before.  If the SNP are back to having a substantial lead over Labour, and indeed a bigger one than they had at the local elections in May and the general election in June, why did Labour do so well in the Fortissat and Cardonald by-elections earlier this month?  I still can't really offer an answer to that question.

Monday, September 25, 2017

The other type of 'winner'

After the result of the German federal election became clear last night, I posted a couple of tweets about the BBC's misleading headline reporting of the subject.

"My standard complaint about BBC reporting of PR elections - you don't 'win' with 33% of the vote. Who governs is decided by negotiation."

"BBC on Merkel with 33% of the vote: 're-elected, wins'. BBC on SNP with 37% of the vote: 'a baaaad night for the SNP, independence is dead'."

Both tweets came in for some strong criticisms, so I'm going to respond here, free from the constraints of 140 characters.  The objection to the second tweet was basically that I was making a bogus comparison, because the two electoral systems are different.  (In fact one person got a bit hysterical and said something along the lines of "we gotta be better than these bull**** comparisons!")  In reality, of course, the difference between proportional representation and first-past-the-post makes the basic point more telling, not less so.  In Germany, the CDU and CSU not only got roughly one-third of the votes, but proportional representation also ensured that translated into roughly one-third of the seats.  Whereas in Scotland in June, the SNP received 37% of the votes, and first-past-the-post converted that into nearly 60% of the seats.  By any measure, then, the SNP's performance was superior to Angela Merkel's, and there is no excuse for failing to report the SNP's result as a "win" if that's what you're going to do for Merkel.

On the first tweet, I got into an abbreviated repeat of an argument I had  with a German chap last year who pointed out to me after a regional election that the German media also invariably refer to this sort of result as a "win" (and, indeed, that the BBC's reporting of German elections seems to take its cue directly from that).  The original discussion went round in circles on the point of whether the German phrase "klarer sieg" means exactly the same thing as "clear victory" in English, or whether there might be a subtle difference that is being lost in translation.  Not being a German speaker I don't know the answer to that question, but put it this way - if the two phrases do have an absolutely identical meaning, the German media are being just as irresponsible as the BBC.  It's reminiscent of what the CBC did in Canada about ten years ago when on election night they "projected a Conservative minority government", which is not something that can even be projected because a minority government is not an election outcome.  The actual result of the election was a parliament of minorities, which is something that might produce a minority government, but might just as easily produce a coalition government which won't necessarily be led by the largest single party.  And yet when the latter scenario started to unfold in Canada, the public thought something was badly amiss because they had already been told that the "result of the election" was entirely different.

The German public are much more used to coalition permutations and are thus unlikely to react in the same way, but nevertheless as a matter of principle it's just wrong to describe something as a "win" when it's nothing of the sort.  The German who took me to task on Twitter made the point that there is no realistic way that Merkel isn't going to form a government, which boosts her right to be described as the "winner".  But, again, that fact has got nothing to do with the election result in and of itself - arithmetically, an SPD-led government excluding the CDU is viable and could even have a huge majority.  It won't happen, but that's because of decisions made by politicians, not by voters in the ballot box.  It's also worth pointing out that German media seem inclined to declare a "winner" even when there is genuine doubt over whether that winner will lead the government.

In an attempt to rescue his argument, the German chap prayed in aid a dictionary definition of "win" in English as meaning "do better than others".  But you only have to think of some hypothetical examples to see how that definition doesn't tally up with how the word is actually used in real life.  If you had a 110 metre hurdle race in which all the athletes hit the hurdles and failed to finish, there would still be one athlete who did "better than the others".  But is he "the winner"?  Of course not.

By the way, just by coincidence, when the SNP first took office in 2007, they did it on the basis of 33% of the vote - exactly the same as Merkel received last night.  If memory serves me right, the BBC did not declare the SNP as the "winners" or say they had been "elected to government".  Instead, it was very carefully reported that the SNP had "broken through, and replaced Labour as the biggest party".