A very good analysis of the road traffic accident data in London is present on this web page: http://www.londonbusroutes.net/miscellaneous/Accident_trends.htm. It has only recently been brought to my attention but it is definitely worth reading if you care about road traffic casualties.
It tackles the question of whether traffic calming, and in particular 20 mph speed limits, have the benefits claimed or are cost effective. One table in the report which I reproduce below is based on an analysis of the accident reductions in different boroughs with differing levels of traffic calming (it gives all the underlying data based on TfL figures if you want to check out your own London borough):
|Level of traffic calming||Reduction for|
As the report says, the correlation is the reverse of that normally claimed. It notes they are statistically significant although correlation does not necessarily imply causation.
The author of the report apparently has a keen interest in buses (most of his web site is about bus routes in the capital), and also says that most of his travel is by rail. So this is not the normal profile of a person who opposes traffic calming – indeed he seems most concerned about the delays to buses from wide area 20 mph schemes. But it is a very intelligent analysis of some of the issues and well worth reading.