Dead Men Left

Monday, May 31, 2004

Polls (plus Latvians and "1.6m gypsies")

I muttered recently and obliquely about the failure of opinion polls to register much in the way of support for Respect. Their reports were wildly, wildly out of line with my experience, and what I hear from the rest of the country: last night, for example, 300 attended a Respect meeting in Tooting; all day today, stalls have been operating across London and if the reception they got was anything like the reception we received on the battle-bus, then we are manifestly having an impact. Having played a part in some Socialist Alliance election campaigns - a generally rather grim experience - with the slightest of slightest successes, the enthusiasm Respect is greeted with is simply off the scale in comparison. (I pity the few poor souls still chugging away in the "Democratic Socialist Alliance". People's Front of Judea/Judean People's Front - splitters!) This is the real "Baghdad bounce", I suppose; the invasion of Iraq has made solid all those vague resentments Blair has built up in his seven years at No 10, and turned them into a substantial social force. Campaigning for Respect just now is starting to feel something like campaigning against the war on Iraq in the build-up to Feb 15: not just polite agreement, but enthusiasm.

So why, then, the absence of this experience in the polls? The most obvious answer is that, despite appearances, Respect is not making a broad enough impact. There is a massive difference between enthusiasm on the stump, and votes turning out come election; and the numbers voting are so large that without a huge electoral machine and access to the mass media, nothing meaningful can be conveyed to them. It need not be that the message is unpopular: many potential votes will disappear through simple ignorance. Respect has been absolutely crippled by the virtual media blackout imposed, and has had to resort to extremely old-fashioned campaigning: relying on community organisations, street meetings, leafletting, and word-of-mouth. We will find out on June 10 whether this is enough to undermine the monolithic structure of the two-and-a-half party system in the barely three months of our existence. The gamble is that the networks established by the anti-war movement, alongside the breach in Labour support that the war has forced open, will be enough to deliver the 5-8% necessary to secure seats.

So it is possible that we have not created the momentum, nor allowed ourselves enough time, to inform enough of the electorate of our existence, let alone persuading them to vote for us. Galloway is a great asset, of course, since - love him or loathe him - everyone knows exactly where he stands, making establishing an identity and promoting a programme far easier; in addition, in a political climate deeply hostile to "establishment" politicians, Galloway - with his cigars, his suits, and his reputation - is clearly anything but, and this is significantly to his favour. Even so, this is a more plausible explanation than that we are failing to communicate an effective message. Aside from sneers coming out of the dustbin of history now inhabited by the fragments of what once passed for a "British left", Respect's message - we opposed the war, we oppose the occupation, we're not of the political class - is a popular one.

More to the point, however, is the flagrant abuse of opinion polls. Though presented as the very pinnacle of objective social research, they are anything but: the statistical guarantee of the central limit theorem means that except in times of significant social upheaval, in which the parameters of the system itself are in question, opinion pollsters can be quite staggeringly sloppy in their methodology, and still produce plausible results. Of course, with the likely collapsed turnout on June 10, and the disengagement from conventional politics amongst those who will vote, we are likely to discover ourselves in a situation of significant structural change. A major base of support remains for the main parties, preventing polling predictions for them straying too far awry, but predicting just how the lesser organisations will perform is likely to be open to quite significant concerns.

Take the UK Independence Party, for instance. Many will have seen the Telegraph poll, suggesting that Kilroy-Silk's happy band of nasty Little Englanders will outpace the Lib Dems come June 10. (With the exception of Nick Griffin, is there a more unpleasant public figure than Robert Kilroy-Silk in Britain at present?) A closer look soon reveals that the 18% support for UKIP is from "those most likely to vote": a hopelessly subjective category, and one that need have no correspondence to those who do actually vote in a few days time. The question on voting intentions was prefaced by a few, carefully-worded questions about Europe - just to get respondents in the moode. Another UKIP-commissioned poll, as reported in the Guardian, was even more overtly biased:

With Robert Kilroy-Silk ready to wrap up the Asian vote in the East Midlands, the Backbencher doubts that UKIP needs any help in the polls. Still, the party was sufficiently alarmed to have commissioned a poll from Mori, which shows it in first place with 35% - yes, 35% - of the vote.

Rather than just asking people how they intended to vote, UKIP took a creative approach. "At the European parliament elections the UK Independence party will be campaigning nationwide for Britain to leave the European Union and put an end to unlimited EU immigration," Mori asked that representative sample of the population that waits at home to talk to pollsters. "Assuming the UK Independence party were the only moderate party campaigning for this, which party would you vote for?"

Yes, it's a new low in opinion polling - and the Backbencher thought the industry couldn't have sunk any further in its efforts to please its clients. Now, what Robert takes "unlimited EU immigration" to mean, the Backbencher isn't quite sure. Could mean they want to come in; could mean they just want to move around. But just to avoid all possible doubt, she is prepared to put in a bid to take that Spanish second home off your hands before journalists start asking questions about it. GBP100? Do I hear GBP150?

This was sufficiently ridiculous as to be unreported outside of gossip columns, but you get the idea. Add to this the small problem that Respect is not offered as a category on polling cards - being left as "other", or requiring the respondent's own initiative - and the net worth of polls in predicting its vote collapses entirely. Opinion polls of elections in which few will vote, and many of those who do will behave "erratically" are going to be (at best) methodologically flawed, at worst open to outright abuse. We are left with nothing but hunches and anecdotes. Respect, I am certain by now, will perform credibly well on June 10. As to how well, unfortunately, I - like everyone else - have no idea.