Saturday, 10 April 2010

Poll Paranoia

Polls during this general election campaign, and before it, for a variety of reasons have been more frequent and less helpful than at any time before. The idea that simply doing more and more polls and then 'weighting' them with a series of arcane, largely untestable methodological tricks, usually hidden from the public (though Mike Smithson, among others, has helped to correct that state of affairs) is one that I think is simply an extremely expensive waste of time carried out by self-styled professional psephological gurus happy to take the cash. The YouGov "daily tracker" is probably the worst offender being, as it is, about as reliable as a Met Office long term weather forecast (see the Daily Telegraph from a few days ago). But it is by no means the only one.

The media loves it, of course, especially if these dodgy measurements of voting intentions suggest a close call - and a hung parliament. Big News! Sells papers and sucks in viewers. So they drive that narrative, using these non-statistics as evidence. The hung parliament trope has become the most popular recently, much to the delight of the MSM, but not, it seems to me, to the Labour party, who have been largely silent on the subject. Why? Certainly not because they are confident of winning. The more-likely reason is that they have about as much faith in these exercises in statistical soothsaying as I do.

Not so the media, however, with a new ICM poll out about key marginals, commissioned by Murdoch rag the News of the Screws and faithfully advertised by Spectator editor Fraser Nelson. Taken at face value, Nelson's argument that the 'extensive research' that took place to produce this poll might seem to make sense; that the Tories are going to need a hell of a lot more than the 5% swing they think they need because, so the poll allegedly shows, they aren't performing as well as they thought, with the LibDums nicking votes from them left, right and centre.

Aside from the fact that Nelson also works for the Screws, and therefore Murdoch, and the fact that it's the parasitic Murdoch media machine, particularly Sky News, that is pushing the hung parliament story at every opportunity (complete with a misleading poll ticker appearing 24 hours a day on their rolling news channel), scratch beneath the surface and you see that this earth shattering poll in nothing of the sort. It's hardly more reliable than a complete guess. Just look at the size of the sample, as mentioned in the small print:
Here's the small print: ICM Research interviewed a random sample of 1001 adults aged 18+ by telephone on 7-8th April 2010. Interviews were conducted across the 96 (new boundary) constituencies which are held by Labour where the Conservatives require a swing of between 4 percent and 10 percent to win the seat.
As commenters on the blog have quickly pointed out, 1001 people over 96 constituencies?! You have to be kidding. That's about ten people per constituency. As I said, they might as well have saved themselves the thousands it cost to do that and just thrown a dart at a bunch of numbers pinned to the wall for all the use a poll like that is. But Nelson still banged on about it as though it was remotely meaningful, proving once more, at least to me, either that he is compromised by his connections to the News of the World, or that he's just not as smart as I thought he was, or that he just as much a fully paid up luvvie member the fourth estate as every other hack and hackette infesting this country, 'blogging' or no 'blogging'.

The point? The point is that my hunch - that the Tories are actually miles ahead, but that people simply aren't ready to admit that yet, or that these polls are so flawed, they would never reflect the reality - is worth just as much, if not more, than the polls themselves. And so is yours, dear reader. So the Tories do not need to develop poll paranoia. They certainly don't need to "panic", as Nelson helpfully suggests in the picture accompanying his dishonest/incompetent article.

The only danger is that all this hung parliament speculation, regardless of the flakey, fake stats that's driving it, will become a self-fulfilling prophesy, thus handing Labour an unexpected lifeline - and selling the country down the river. No wonder the Tories kept Murdoch's empire at arm's length when it turned coat again and came out for them. With friends like these. In addition, it's plausible that poll paranoia will catch hold in the Tory party itself, again making the whole thing self-fulfilling.

Nevertheless, despite the media's attempts to drive the outcome of this election, I firmly believe that the damage that's been done to Brown Labour, and that it's done to itself (not least by allowing Brown to become leader unopposed in the first place) is so grave, not even corrupt media organisations and shockingly crap pollsters can save them now.

Have I gone too far? I'll tell you on May 7th. Or someone will tell me, no doubt. What I suspect, though, is that on May 7th, Britain will have a Tory government with a comfortable, if not large, majority - and an awful lot of pollsters and tame journos, once the noise from the celebrations have died down, are going to have an awful lot of egg on their collective mugs.


  1. I agree. From my own, extremely scientific, poll of one voter (me) I can predict a huge Tory majority!

    Seriously, my guess is a 50 seat majority.

    I do think all this stuff about a bigger swing in the marginals is nonsense though. I reckon we'll get a uniform swing big enough to win.

  2. I think I need to learn to be more succinct ;)

  3. I reckon the Tories will win with a majority of at least 80, D, the current blizzard of polls are about as useful as a one-legged man in an arse-kicking contest! I think you can safely add at least 7% to the Tories' share of the vote and I totally agree with everything you said in the post - I've had the MSM and rolling news channels right up to here!


Any thoughts?