g_attrill wrote:
Paul, I have just acquired the Road Accident database 1991-2002 and am currently playing around with some figures for some Portsmouth cameras.
I am seeing some very interesting results, for instance the Mile End Road camera (right at the start of the M275 on a six-lane mega junction) the KSI has been a couple a year for the majority of the 11 years but was 4 in 2001, this would have skewed the 1999-2001 KSI total which is 6 in their latest spreadsheet (
www.safetycamera.org.uk). 2002 had 1 KSI. This means that if 2003 was like the other 10 years (0-3 KSI) they would be able to claim a 50% reduction.
Gareth
Excellent. Can you (we?) perform this exercise with either: ALL cameras in an area or a random sample of cameras in an area? We have to be most careful not to cherry-pick sites that are favourable to our arguments. It must be a large group or a random group.
We also have to be careful about site sizes - if the defintion of the site changes because of thoughts of cameras this might skew the figures.