The "evidence" in favour of cameras has always been dubious. They lump virtually everything under "speed related".
This is the text of an article in Motor Cycle News July 3 2002
"Take a look at the Transport Research Laboratory report TRL323, which lists accident causes according to police officers at the scene. It's this document which is behind the DLTR's claims that one third of all accidents are speed related, the prime justification for the plague of speed cameras.
The single biggest proportion of that "one third" comes from 'failure to judge another person's path or speed' at 10.7 per cent of ALL accidents. This is astonishing -- it means that when someone pulls out of a junction in front of you and has you off your bike, you have just statistically become a speed related accident used to justify more speed cameras, regardless of what speed either of you were doing!
There are lots more examples: "following too close" is, according to the Government, speed related too (even if the driver is following too close at 50mph on a a motorway with a 70mph limit, because the driver in front doesn't understand the 'keep left unless overtaking' rule).
Slippery roads are also counted as speed related by the DLTR, something to mull over as you slide off on a roundabout doused in bus diesel at 10 mph. So is bad weather, even though going too fast in bad weather is listed separately under excessive speed. Riding slowly in the winter but coming a cropper on unsuspected black ice is, for the DLTR's purposes, still a speed related accident fuelling the need for more speed cameras.
As for "excessive speed" itself -- TRL323 blames this for just 7.3 per cent of accidents. And that includes excessive speeds which might be below the posted speed limits, which speed cameras do nothing to address."
Eventually the government and camera partnerships (mostly) stopped using the one-third figure, but they still seem to be stuck on the idea that most accidents are caused by speeding.