January 09, 2007

If I wanted to know my opinion

Roy Morgan rang me last night to find out about my views on "topical issues". It turned out the topical issue was buying tobacco products, and my views were sought in the form of whether I'd bought any. Perhaps I'm a bit precious in this regard, but I think my views extend beyond a yes/no to buying pipe tobacco (no, as a matter of fact).

Roy went on to ask me if I'd be interested in signing up for their special opinion survey by email service, in which the service is that I get to "have my say" and receive a "special benefit" (no, as a matter of fact). I never found out about the benefit, but the idea that opinion surveys are a form of having a say had me intrigued. Having a say in what? What sort of a say? Having a say to whom?

Given most politics is done by polling these days ("we are in favour of ") they have a point. Opinionative Democracy - The Likert Scale of Political Aspiration. It could be a big new philosophy. We could start to characterise political parties by how they like to see opinions cluster on the scale. The Libs and the right wing of the ALP like to get a nice solid grouping of 3s; the left wing of the ALP and the Greens will sacrifice 3s and live with 2s as long as they get increased 4s. Branch stacking will become mostly a matter of buying enough people to give 5s on the internal party polling.

In fact, I propose that Parliament be reformed along opinion polling lines. Instead of electing members to seats, we should offer opinions on candidates. The party with the most 5s ("most representative of the things I like and/or buy") gets government, whilst a set number of seats is reserved for the party with the most 2s ("not very representative of the things I like and/or buy") to represent the shadow market (as described by David Foster Wallace).

Elections will be still be overseen by the AEC, but will be conducted by competing bidders from the market research vendors through a series of random telephone surveys. Hang on, you're saying: random surveys? What about everyone getting a vote? But that's the beauty of the idea. Your opinion isn't going to be very different from anyone else's, and even if it is, the physical constraints imposed by product supply mean that you can't do anything with it (you try voting for Yo La Tengo in the next election and see where that gets you). So random surveys are perfect. They're statistically reliable enough to be as representative as any other method. The Likert scale is just as precise as the preferential system. The question isn't very different - you don't even really get a question in traditional elections, you just get to put a number in the box corresponding to the implied question of who you want to send to Canberra/Spring St/Gower St/hell.

And the true golden joy of the idea is that it completely short-circuits all the tedious debates about compulsory voting and whether it's democratic. Just poll people. If they don't want to give an opinion, they don't have to. No one's making anyone do anything, but at the same time if you want to have a say, you can, and because the market researchers have a set number of answers they need to get statistically reliable data, they can just keep phoning until they've got enough responses.

posted on January 09, 2007 at 02:00 PM by darren.