Thanks for the tip-off 'anon'.
Just been sent this link:
http://www.saferroads.org/news/latest-news.htmlWhich points to this newly created document:
http://www.saferroads.org/media/pdfs/ca ... port09.pdfFirstly, it completely blows out claims that accidents are reduced by 40-60% at camera sites - it's not even 20% (3 year average 'after period' compared to 3 year average of baseline). So in the Thames Valley case there isn't much RTTM because there isn't much in the way of reductions anyway - sneaky!
Even worse, the effect of any inflated baseline
for each installation is averaged out by the other installations which have their own inflated baselines
at different times. This is mathematical masking of the naughty variety.
Of course none of that begins to address the issue of 'bias on selection' (other safety measures applied within camera sites), which would lead to genuine, permanent reductions at the camera site and would easily explain the difference between the two trends.
But then I realised. The graph which the conclusion is based on doesn’t show KSI casualties, it shows only collisions. Speed camera placement criteria isn’t dependent on the number of collisions; it's not the standard measure of their success either .......