I tried a little statisitical experiment for my kids maths homework last week:

I tossed a coin a hundred times and got sixty heads.

Amazing: it should have been only fifty!

So I tried it again after setting up my video camera to record the event.

And only got thirty heads this week, instead of sixty!

Clearly my video camera has reduced the chance of a head by 50%!

 

Next week I'm going to set it up at the scene of a road accident.

And see if I can reduce the chance of a head on collision by 50%!

 

So, when your KSI figure increases from 1109 in 1997 to 1536 in 2003 (after installing speed cameras) it does not matter as long as you show that you have halved the number of accidents that occur within 5 metres of where you have put the cameras (i.e from, say, 2 down to 1 accident).

Well, that's what Thames Valley seem to think.....

"We welcome the report and whilst we are concerned about the overall (INCREASE IN) casualty levels we must not lose sight of the fact that we entered the pilot project with a history of almost 10 years use of camera enforcement which had already resulted in significant reductions in Death and Serious Injury.

Our own figures show that we are continuing to affect the most serious road crashes especially at our targeted sites. We are also having a much wider effect on traffic speeds which not only affects casualty reduction but also improves the environment for our communities. We regret the fact that the report fails to fully recognise our previous contribution and that the base line for the study was set after those reductions had occured......

....Whilst the report has excluded Thames Valley’s KSI performance (because of the changes in reporting procedures which we introduced in early 1999 to improve compliance with Government statistical requirements), our own local results do show reductions in Killed and Seriously injured casualties and collisions particularly at our targeted camera sites."

http://www.saferroads.org/DisplayPage.asp?pageid=18

Or download the file: http://www.saferroads.org/Downloadfile.asp?file=16_10200319020313_75.doc

PROBABILITY SHOWS ROADS ARE NOW SAFER

10:30 - 06 May 2004

Your headline "One in three speeds cameras failed to cut accidents" (April 17) does not, as is implied, mean that speed cameras are not effective.

In order to understand this, think first of the situation in which a coin is tossed, say, 10 times.

On "average" you would expect five heads and five tails, but in practice you might get four heads and six tails, or seven heads and three tails and so on.

If you repeated the experiment lots of times, you would expect to get more heads than tails about the same number of times as you got more tails than heads.

This is a bit like the speed cameras. If speed cameras have no effect, then you would expect the number of the locations where the accident rate increases to be roughly equal to the number of locations where the accident rate decreases.

In fact, the accident rate increased at 21 locations, but decreased at 60 locations. This suggests cameras are very effective.

It is a shame to see the press making an effort to discredit speed cameras instead of applauding their contribution to cutting accident rates and encouraging obedience to the law.

http://www.thisisleicestershire.co.uk/displayNode.jsp?nodeId=132938&command=displayContent&sourceNode=132547&contentPK=9852449

HUH?

>In fact, the accident rate increased at 21 locations, but decreased at 60 locations. This suggests cameras are very effective.

Ah, but, when the accident rate increases at 60 locations, but decreases at 21 locations.
This suggests cameras are very ineffective?

Nutty Professor.
:-)