Thursday, April 19, 2007

How you will vote

There’s a very interesting analysis of voting in the last Federal election, 2004, courtesy of the Oz Politics Blog website.

The researchers (Stimson, Dr T- K Shyy, and Dr Prem Chhetri) analysed each polling booth, measuring who won the most votes, and who won more than 20% of the total for the booth.

Matching this with the 2001 census, they analysed 46 variables, and were able to explain 96% of the variability between booths, on the basis of

Multicultural/younger vs monocultural/older; and
Advantaged vs disadvantaged.

The simple result:
Multicultural/younger and disadvantaged -> voted Labor
Multicultural/younger and advantaged -> voted Green
Monocultural/older and advantaged -> voted Liberal
Monocultural/older and disadvantaged -> voted National

It’s been charted along those two axes. Of those four, the National’s point in Monocultural/older and disadvantaged was furthest away from the point of origin, suggesting the strongest affinity with that characterisation. The weakest characterisation was for the Liberals.

(It would be interesting to see how it changes at the forthcoming election, since it looks like the current government is headed for a landslide loss. It’s possible the variability doesn’t change, so much as the base vote.)

Have a look at the Oz Politics Blog. The picture makes it clearer.

All this is in the aggregate. So it's not deterministic on an individual level. But it will tell you something about the people around you.

Other political sites I've found interesting include:
Larvatus Prodeo
The World's smallest political quiz
Poll Bludger
The Road to Surfdom
Pavlov's Cat

But you can't beat the banter in the comments on the Oz Politics Blog. Often some useful insights - with a little patience. The Oz Politics site itself also has a good top-down overview of Australia's political and voting system.

UPDATE 20-Apr-2007: Two very interesting comments at the end of that Blog entry are well worth noting.l

First, Anthony Green [Australia's foremost electoral analyst] noted that voting patterns were not, by and large, predicated on age, but on the milieu in which their voting habits emerged. So people voting for the first time in the 1970s would display rather different voting patterns - throughout their lives - than those first voting in the 1950s.

Second, a problem in taking results in the aggregate was illustrated by a finding that in the US, wealthier individuals voted Republican, but wealthier States voted Democrat. Referred to as the "ecological problem", I think? Fascinating, and a principle [of caution] worth remembering.

(If blogs are the obscure minutae, then blog comments are the hidden quarks in the atom...)

No comments: