Safe Speed Forums

This might be very interesting.

I suddenly realised that the ratio of Personal Injury Collisions (PICs) to Killed and seriously injurded accients in the camera placement guidelines is 2:1, but in the real world PICs:KSIs is 6.66:1. I realised that this means that PICs are unlikely to form part of the selection criteria for the average camera site - they are too plentyful by a factor of three.

So I went the the DfT's new spin report and found:

40% fewer KSI 870 fewer KSI
33% fewer PICs 4030 fewer pics

This equates to 1450 before KSIs and 6015 before PICs

The percentage of PICs which were KSIs before the cameras was 24.1%
The percentage of PICs which were KSIs after the cameras was 21.6%

From the DfT's "2003 main results" we find 32,160 KSI accidents and 214,030 PICs. The national percentage of PICs that were KSIs was 15.0%

Now, is the difference between 15.0% and 24.1% the selection bias that causes the regression to the mean benefit illusion or is it something else?

15% is about 40% less than 24%. So if these sites return to normal ratios in KSI we'll get our 40% RTTM benefit illusion.

We can also do these sums on a "by county" basis and maybe build up more of a picture. I don't yet know what I've found this time, but I'm very interested indeed in it.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety