The TV debate: Colourless green ideas sleep furiously
I’m not over-fond of quoting Chomsky, but on this occasion his most famous sentence is the key to the television debate last night.
It is possible to torture a meaning from ‘colourless green ideas sleep furiously’, which may apply to the debate. Natalie Bennett was competent, but didn’t light up the stage, sure. But what ‘colourless green ideas’ conveys is that it’s possible to construct a sentence that is well-formed but completely meaningless, because it’s a category mistake on compound levels. It’s self contradictory, twice, and attributes properties and capabilities to something that cannot possess these qualities.
Televised leaders’ debates illustrate that it is possible to ask a question, conduct a poll, get an answer and spend an inordinate amount of time discussing a meaningless category mistake of an issue.
Who ‘won’ the debate?
We might as well ask which of the paintings therein ‘won’ Tate Britain. It’s a silly question.
For a start, who is the relevant audience? If you confine the polling to those who saw the broadcast, that’s good because they have a chance of knowing what they are talking about. But it’s also unrepresentative of the general population – they are likely to be more political and more committed. Do you try to expose a representative sample of the population to the debate and see what they think? Whichever you do, the chances are that people who supported a party before the debate will tend to think their man or woman ‘won’ the debate even if you ask them to ‘leave aside your own political preferences’ as YouGov did. People can’t easily do this because perceptions are biased. You might as well ask fans at a football match to put aside their team preference and tell you whether that really was a penalty or not.
Do you try to find a group of undecided voters, and see how many of them decide their vote at the end and proclaim a winner? Or weight to a politically representative sample of the population somehow? What is the relevant baseline? Voters may decide this in varying ways. Do we measure from a baseline assuming zero knowledge or equal esteem for each candidate? Then the chances are that the candidate who was ahead at the start will ‘win’. Or do we measure change, which will probably favour the previously unknown candidates? Or do we try to factor in a ‘par’ in which someone who over-performs low expectations has ‘won’? To what extent does one try to prescribe how people answer the question? Should you calculate a net score by taking away all the people who thought that party leader was worst or whose opinion of him or her fell? All these choices will bias the coronation of a ‘winner’.
The debate polls are all over the place, unsurprisingly given the absurdity of the question and, as Andrew Sparrow significantly points out, the effects of different ways of weighting the samples. General election polling has an objective measurement with which the performance of the polls can be verified – to wit, results of said election. If your methodology is wrong, real life usually reminds you of the fact – you predict an election inaccurately, or the advertising approach you recommend doesn’t increase sales. There is no objective measurement for debate performance.
And what on earth does ‘performing best’ mean? Ever since public debates like this started with Nixon and Kennedy in 1960, we have known that the arguments are a bit secondary. Famously, viewers on television thought Kennedy had won while the radio listeners gave it to Nixon. Are we deciding who looks better, who looks more confident, fluent and assured, who has the best jokes? If one starts to think about ‘being Prime Ministerial’ the advantage tends to be with the person who has been in the role for the last five years. But it’s probably mostly about fluency, confidence and entertainment value.
It might be just about possible, with lots of hedging, to declare a winner in a head to head debate between two people. But doing so in a seven-way discussion should lead us to have a bit more of a think about whether someone who heads the poll with a 28 per cent rating, let alone 21 per cent, has ‘won’. Winner-take-all electoral systems give weird results when you have lots of options. Very few people (other than the most determined partisans) will have formed the view that their chosen ‘winner’ was excellent and everyone else on the panel was rubbish. The Guardian/ ICM poll did at least do a forced choice question between Miliband and Cameron (the answer was 50/50 for what it’s worth).
I’m Sorry, I Haven’t a Clue
A lot of debate analysis, about the strategies and feints of the candidates and who benefits, reminds me of nothing more than the bits of the game Mornington Crescent where they start to talk about its ‘rules’: “Ah… clever. The Northern Line parallel is, I suppose, allowed under the post-1965 Rapid Play Rules. But not clever enough, Beck’s amendment means I can do the counter-switch at Euston – Mornington Crescent!” “Miliband did need Sturgeon to do better than Bennett, but if you look at the cross-tabs the ABC1s didn’t think much of Leanne Wood so you have to discount that down to 15 per cent – David Cameron obviously won.” Some sections of the press (and one could have expected better from the Telegraph) had sufficient contempt for their readers not to bother even with dishonest or over-sophisticated arguments, and just trundled out the stories they had already written bashing Ed Miliband.
The worst way of deciding who ‘won’ the debate (if I haven’t convinced you not to bother) is reading biased newspaper headlines. The next worst is to read off the top line of an opinion poll (and you can take comfort that there’s probably one poll whose details can support whatever prior views you had about what counts as a winner).
I have a lot of time for the partisan who bothered to watch the debate and is genuinely convinced that their man or woman was best – it’s generally an honest response even if affected by perception filters and bias, just like a post-football match chat in the pub. If it heartens activists to go out and talk to more voters and deliver leaflets, that’s fine with me. If some voters have taken the art gallery approach and have an appraising look at all the options, even better.
Jamie McDonald’s words to the hapless Cliff Lawton in The Thick of It were wise indeed. The leaders aren’t horses, with or without expletives, and it’s stupid to talk as if they were. Seeing a seven-person television debate as a horse race – and a particularly Alice-in-Wonderland style horse race where there’s no agreed finish line, the volume of cheers from the audience is taken into account and all the runners take it in turns to stop for a bit – is just as ridiculous as saying that a door handle hates you or that you saw the King of France driving a train on the Victoria Line. But can we really say that the whole meta-, horse-race, game-playing discourse we use when deciding some quite important things is any more rational?