Sunday, 2 January 2011

Do we make them justify anything?

In case you haven't worked this out already: I really, really don't think much of the road safety discussion currently being played out in the state and national media.

It's important but it's crude, simplistic, over-wrought and driven far too much by government and police sound bites, and I don't even mean the almost exclusive focus on deaths to the exclusion of permanent trauma causing lifelong disabilities at a tremendous emotional, physical and financial cost to individuals and their families.

And most of the discussion orbits the topic of speed cameras.

In order to justify the use of speed cameras as a road safety initiative, not merely a fundraising initiative, there needs to be a direct and confirmed causal link between the usage of cameras and an overall reduction in the rate and severity of accidents. Not just a correlation - a causal link. There are too many possible factors impacting driving behaviour for us to jump to a simplistic conclusion if there's a dip in the death rate.

Then, if we've done that, we need to be sure that there are more effective, more cost-effective ways of addressing the road toll. Maybe speed cameras work, but not as well as some of the alternatives.

Maybe the road and traffic authorities know what they're doing. But it's really hard to be sure of that, because they're not talking as if they do.

Let's start with More speed cameras deployed to cut road deaths from the ABC (December 15).

It tells us: Police are trying to cut road deaths; they are putting camera vans in 40 and 50 kilometre per hour zones; there are more camera vans in service (12, now); in two years 57 people have died in 40 and 50 zones; the road toll is lower than it was last year, and; putting camera vans in low-speed areas is part of a plan to cut the road toll.

There are logical holes in that. There is no justification that camera vans help to lower the road toll, they are concentrating resources on areas with a very low percentage of the toll (57 in two years, when the two-year total toll is in the region of 600) and where death is less likely due to the low speeds anyway.

Are non-suburban areas less important? Or is it harder to put resources there, so it's too much of an effort?

There is not even any indication of how many deaths are related to speed anyway, which would give the public some hope of working out for themselves if the tax money they're paying to be used in crude behaviour modification on them has any hope of being spent in a way remotely resembling good value.

From the Brisbane Times (December 18), we have Sixteen reasons for covert speed cameras.

Ooh! Justification!

Hahahahaha, psych! No, it's just another article saying "We have road deaths, let's do something simple and easy."

Here's the interesting part, right up front: 16 speed-related deaths in two years in 40 and 50km/h zones. That's 16 out of 57 in low-speed zones, or 28 per cent, or less than one third, and that's out of a tiny proportion of the total. Tell me: What are the police doing about the 72 per cent of deaths that weren't speed-related? Anything?

If we take a two-year total of about 600, that is, as pointed out in the comments to this article, about 2.7 per cent of deaths getting the added attention of two covert camera vans. May I suggest that's a little disproportionate?

There is, however, a balancing although inadequately explored opinion delivered by National Motorists Association of Australia, saying that covert cameras are all about revenue raising and can never be as effective as police on the streets.

It would be nice to have some sort of discussion about why that is.

Finally, appearing in the RSS feeds as "Speed cams are lifesavers" (what, muscular men in red speedos?), the Courier Mail gave us Covert speed cameras linked to Queensland plummeting road toll.

Okaaaay. Linked by whom? Statisticians? Accident researchers?

Well, I've read the entire article twice and I'm still not sure. It's not really stated.

The salient points are: A poll showed 25 per cent of motorists have changed their behaviour as a result of covert speed cameras; quite a lot of money has been charged in fines this year, and; there are concerns of a bounce in the accident rate as people adjust their expectations and return to previous behaviours.

The whole "cameras = less deaths" thing is only ever hinted at.

It is assumed that "changed their behaviour" means "safer." They might, instead, be driving slower but angrier and more aggressively. They may be safer but it could be because people are more alert on the roads trying to spot covert police and are therefore more aware of road and traffic conditions, rather than because they're driving slower.

If that's so, then good. But there might be better, more generally useful ways of getting the same result.

It's even assumed the current low death toll has any explanation beyond annual fluctuations. I've forgotten too much about statistics to know what the thresholds are and gee, it'd be nice to hear someone explaining it.

As an exercise, I went to PubMed, typed in "speed cameras" (with quotes, to get the actual phrase) and had a quick read through the abstracts of relevant articles.

There's the usual problem of identifying articles that have anything to do with the topic, and then when you find them there are a nice smattering of methodological problems. I even found one article finding a difference in effectiveness between different types of roads, and then another article using one category of road as a control and the other as an intervention group.

There was also, astonishingly, a Cochrane review, conducted by researchers right here in Brisbane.

Going to the source, we find that the Cochrane Library contains 12 articles relating to prevention of traffic crashes.

(NB: The conclusion of Motorcycle rider training for the prevention of road traffic crashes is "We have no idea, no study has been of a high enough quality to justify drawing any worthwhile conclusion.")

The camera review is here.

A review of 35 relevant studies concluded that cameras are associated with lower speeds and lower rates of accidents, but the degree of effectiveness can't be confirmed because the studies are of varying and generally moderate quality.

There, see? We have a result: Research has concluded that speed cameras help lower the road toll. It took me five minutes to find that.

Now, before I have any faith that my tax dollars are being well spent, I want to see:

  • Discussion on the best form of cameras - point-to-point, visible fixed, visible temporary, covert temporary, and visible or covert vehicles in the traffic flow.
  • Discussion on what to do about all the accidents that aren't related to speed.
  • Some sort of visible commitment to research-based policy.
  • Research into why cameras have that effect, and how to maximise it.

I know, "tell him he's dreaming'."

We can hope.

2 comments:

Kit said...

I drive through the CBD streets late at night (mostly empty) to collect J from the movies (Kingston to North Hbt) once a week. I've become accustomed to the traffic lights being in sync for most of the trip... and so I drive exactly as fast as I can to make every light. But last week ...my god - they'd slowed down the 'driving in sync' speed (perhaps just for the holiday season?). Then I realised they'd increased the delay between the red light in one direction and the green in the other. At first it irritated me. But then I thought about the implications for road safety, and now I think it's a bloody brilliant idea. Of course, once the 'holiday death toll count' season is over, they'll likely revert to the old system. Just like the crap we get on TV during the 'non-ratings' months).

Jonathan Hepburn said...

@Kit: If that's really why they did that, it's an almost startlingly intelligent idea. I am always more in favour of carrot approaches that give people incentive to adjust their own behaviour. It will, of course, still annoy people not intelligent enough to think about it.

Search This Blog