http://www.guardian.co.uk/Columnists/Co ... 70,00.html
The anti-speed-camera campaign is built on twisted truth and junk science
Petrolheads are full of swagger in attacking road safety measures, but can't cope when called to account for skewing data
George Monbiot
Tuesday November 13, 2007
The Guardian
Loth as I am to threaten my reputation as a bilious old git, I feel compelled to shock you. I am going to praise the government. New Labour has done something brave. Last week Jim Fitzpatrick, the transport minister, said he intended to double the penalty for drivers who break the speed limit by a wide margin. This means that people could lose their licence after committing just two offences. The newspapers are furious. The petrolheads have called for a petition that "will get as big a response as the road-pricing one".
Well, yes it's brave, but not quite as brave as you might think. Despite endless attempts by the media to trivialise it, an RAC survey reveals that 62% of drivers still regard speeding as a serious offence. Even more surprisingly, as many as 82% of British people surveyed approve of speed cameras, and that percentage has risen slightly since the mid-1990s. There is a genuine silent majority here, which is rarely represented in the media.
Nowhere is more nonsense spoken about this issue than on the BBC. Its Top Gear series has become a sort of looking-glass Crimewatch in which the presenters enlist the public to help criminals foil the police. There are tips on how to avoid prosecution and endless suggestions that speed cameras are useless or counter-productive. The tone was set in 2002 when the team demonstrated that you could beat the cameras by driving past them at 170mph. Since Richard Hammond's crash last year it has had to temper the message a little - but only a little. How, while BBC editors are sacked for misnaming the Blue Peter cat, does Top Gear remain on air?
In the Sun, Top Gear's lead presenter, Jeremy Clarkson, abandons the wink and the nudge for blatant incitement. "As I drove down the M20 into Kent last Monday, I noticed that most of the speed cameras had been burnt out by vandals. This is disgusting. It is ridiculous, criminal and stupid that the person who savaged these life-saving devices should target the M20 ... and then stop. Why did you not keep right on going? I can think of six cameras on my way home that would be immeasurably improved with a spot of petrol and a match."
The tabloids throw up their hands in horror at every other species of crime. They praise the police and demand that the forces of law be given greater powers and that lawbreakers serve longer sentences. But on this issue alone, the tabloids take the opposite position. Richard Brunstrom, the North Wales police chief who is waging war on speeders, is denounced by the Daily Mail as the "mad mullah of the traffic Taliban". The Sun calls him "barmy" and "a politically correct prat". So much for their demands for zero tolerance.
In Saturday's Telegraph, Christopher Booker and Richard North published a long article appropriately titled "Speed cameras: the twisted truth". A sharp decline in the death rate on the roads suddenly slowed down in the mid-1990s. They attribute this to the government's new focus on enforcing the speed limits, especially by erecting speed cameras. What they fail to mention is that deaths started falling sharply again in 2003, after the number of speed cameras had doubled in three years.
They used similarly selective data to argue that there is no evidence that cameras have reduced deaths even at the spots where they are deployed. They hang their case on an oversight in a government report published in 2003. The report claimed that the accident rate had fallen by 35% where cameras had been installed. Booker and North rightly observe that it had failed to account for a statistical effect called "regression to the mean". There might have been an abnormal blip in the accident figures, which would have returned to background levels of their own accord. The truth, they maintain, is that "speed cameras actually increased" the rate of accidents.
But what Booker and North fail to tell their readers was that, in 2005, the government conducted a new analysis that took account of regression to the mean. The fresh figures showed an average reduction of 19% for collisions that caused deaths or injuries after speed cameras had been installed. Why do Booker and North fail to tell their readers that the statistics had been corrected and still showed a major decline in the number of accidents?
Their article is a long catalogue of intellectual dishonesty. In support of their claims that speed cameras are worse than useless, they also use a report by the House of Commons transport committee. It said, they maintain, that "an obsession with cameras was responsible for a 'deplorable' drop in the number of officers patrolling Britain's roads". It says nothing of the kind, and the word "deplorable" does not feature anywhere. But here's what it does contain: "Well-placed cameras bring tremendous safety benefits at excellent cost-benefit ratios. A more cost-effective measure for reducing speeds and casualties has yet to be introduced." Booker and North also lay into one of my columns. That's fair enough: it's a national sport. But to make their narrative more convincing they alter the date of the column by a year. Their claims about speed cameras, like much of the material in their new book, are pure junk science, cherrypicking the helpful results and ignoring the inconvenient ones.
All these people turn, as a final resort, to a man by the name of Paul Smith, who runs a campaign called Safe Speed. He's quoted whenever there is a speeding story in the news. He claims to have found, through statistical analysis, that "speed cameras make our roads more dangerous".
In 2005, he challenged me to a radio debate. I accepted, and floored him with a simple question. Has he published his analysis in a peer-reviewed journal? A peer-reviewed journal subjects new scientific claims to expert scrutiny. Without it, those of us who aren't experts can't tell whether claims are a work of genius or total hogwash. No, he hadn't. In fact he had been asked by the leading journal in the field (Accident Analysis & Prevention) to submit his work for review, but he hadn't taken up the offer as he didn't "have time". (He went on to boast that he had spent 10,000 hours compiling his website.) But he said he would seek to publish a peer-reviewed paper within six months.
I rang him on Friday to ask how he was getting on. "I never did see peer review as a particular need," he answered. "I mean for heaven's sake, there's so much peer-reviewed crap out there that it's just not a modus operandi for us." So just what is the status of his evidence? Beside the statements on his website, Smith lists "source, justification and links". His central claim is as follows: "We simply don't believe that a significant proportion of accidents are caused by exceeding the speed limit." If he cannot demonstrate that this is true, his entire case collapses. Its source, justification and links? He cites this and only this: "Pure opinion, based on considerable driving experience."