Lies, damn lies, and targets

Last week the Scottish government released the latest data on serious injuries on Scotland's roads.   The headline on the Ministerial statement that accompanied the data release claimed that "Scottish road casualty statistics moving in right direction" on the basis that "the number of fatalities on Scotland’s roads are down 8% since 2011, with the total number of road casualties down by 2% to the lowest figure ever recorded" Oddly, quite a lot of people didn't agree with him, because if you looked a bit more closely at the data, it became clear that while fewer motorists were being killed or seriously injured, the numbers were going in the opposite direction for cyclists, and more pedestrians were also killed.  The Minister's statement in fact had no numbers at all for vulnerable road users, while the figures in the data release it accompanied highlighted that there were 9% more casualties in 2012 than 2011.  

However,  the data is much worse than this.  In comparison to their baseline data from 2004-2008, there's been a 17% rise in cyclists killed or seriously injured (KSI) on built-up roads and 34% on rural roads.  

Over at Pedal on Parliament, we pointed out that this data didn't seem to back up his claim the week before, that there was no need to strict liability, because accident rates were already dropping. 

But there is a broader issue at stake.  The government has 'targets' for making roads safer.  They measure these by looking at  'people killed or seriously injured' and 'children killed or seriously injured, and deep in some excel files, they work out what the numbers would be if they met their targets.  The the 'target' for how many cyclists would be seriously injured is 92, but in fact there were 167.  Meaning that they missed their target by 81%.*  

I'm not convinced that we should have government by 'targets', but if we're going to set targets, and collect data, surely we should be looking at targets for vulnerable road users separately from motorists?  Not just because the issues are different but because keeping that data at the aggregate level masks what's really going on.  The statistics don't lie, it just depends on how you look at them, and how you present them. 

If rates for some road users are going up and others are going down -- more than can be explained by increased rates of cycling -- then we need to ask why.   

The only reason for setting targets is so that policy can be designed to address the specific issues needed to make a change.  So, if we had a target to bring down the rates of pedestrian and cyclist KSIs, we'd need to think what interventions could achieve this most effectively. 

And that's a discussion we really need to have.  

*corrected 1/7/2013 after error pointed out via twitter.