Useless statistical indicator alert
Russell Degnan

"The problem with employment statistics is that they encourage us to regard people as potatoes"
- Carlo M. Cipolla, Before the Industrial Revolution

Cipolla was referring to development in the above quote, but it applies equally to almost all statistics in some way or another. It applies doubly to The Age's decidedly questionable liveability indicators.

I have three general issues with this survey.

Firstly, the problem correctly identified in the article that housing preferences vary markedly between people, and therefore, the weight given to different aspects of liveability is only meaningful for the person doing it. A table that allows someone to say "South Yarra is Melbourne's most liveable suburb" is bullshit, much like Melbourne's "liveable city" tag.

Secondly, in addition to the few individual issues to be discussed directly, there is a whole category that seems to have been lost. Nowhere is private land mentioned, in housing size, and lot size, and all those things that really don't change very much, and certainly no more than aspects like tree-cover, culture or crime. It is a huge reason for people choosing not only the outer suburbs but regional and semi-rural areas to live in. Enough to bring the entire survey into question.

Thirdly, the numbers are... interesting.

Take travel, encapsulated in four figures: CBD proximity, train, tram and bus availability. Quality is important here, North Melbourne gets a low bus score but has one of the few buses that actually runs semi-frequently. Similarly, it gets a high station score but has the least accessible train station of any suburb. But moreover, what matters is not whether there is transport but where (and how quickly) it can take you. Why, for instance, doesn't proximity to a freeway count - do the surveyors think people don't value that? And surely the fact that South Yarra station (4) goes many many places is worth more qualitatively than Oak Park (5) that goes barely anywhere? Not to mention walking, cycling...

Geographic values -- coastal proximity, topographical interest and tree cover -- are equally bizarre. I don't know how they calculated topographical data but it looks a bit random -- take the values in the rolling hills of the city and compare them to flat Carlton or Collingwood. Coastal proximity is important, but what about rivers, and creeks? Tree cover is important, but naive if it doesn't distinguish between private backyards and public tree-lined streets. The urban environment seems to have got no status at all, and yet, if one examines the pictures of Melbourne's change -- hopelessly CBD centric as it is -- it can be neatly summarised as: trees, cafes and an active street front.

Why measure shopping (giving the Carrum Downs safeway a 5) and not shopping quality that would highly favour places near to the major centres? Same with schools. And the same with open space, where sports and recreation activities and the quality of the space -- which is much higher in inner and middle suburbs despite being relatively smaller -- should count for far more.

And traffic congestion. Always a nice measure to make people in the outer suburbs think they've got the good life. But congested traffic affects outside residents much more, because they have to push through it. People who live in the inner suburbs, who can walk most places and have often blocked off their residential streets making them quite quiet, have far fewer congestion issues than the statistics might imply.

Using funky GIS data to generate a ranking is cute, but it is like owning a mountain of gold on Mars. You sound rich, you look rich, but you can't do diddly-squat with it.

Sterner Matters 20th August, 2005 13:47:32   [#] 


The Age demographic
I wonder what the correlation between the proportion of people who regularly read The Age and the rankings of the suburbs is?

I have a sneaking suspicion it's rather high...

Oh, and as for the mountain of gold on Mars, you just need patience :)
Rob  22nd August, 2005 14:48:15