This project is being carried out jointly with Kelly Mara for the Land Transport Safety Authority. In recent years, speed cameras and compulsory breath testing have been introduced as road safety measures. Also there has been a major reorganisation of the traffic policing system (more than one?). The object is to see if we can identify any change in accident rates due to the safety measures or the reorganisation.
Over the last few years there has been a gradual decline in serious and fatal accidents. In addition there appears to be an increasing traffic flow, at least on rural roads, possibly accelerating over the last year due to the economic upturn. Deregulation during the 1980s has lead to the import of low cost second hand imported cars and this has lead to a major reduction in the use of motorbikes and possibly of older cars; but possibly an increase in the access to cars.
Good records are available on the fatal accidents. Probably about 50% of serious injury accidents are reported. Reporting rate on minor injury accidents is very low and we are not using this data. Detailed records on the use of speed cameras are available but there is less data on the effort being put into compulsory breath testing. We do not know of consistent detailed information on other safety campaigns. We have reasonable information on trends in traffic flow on rural roads; similar information does not seem to exist for urban roads.
Somehow we have to separate the effect of the speed cameras and the compulsory breath testing from everything else.
The main methods we are using include graphing a variety of breakdowns of the data, cusum charts and Poisson regression models. We have been using S-plus, MS Access and MS SQL Server on our 486s (now Pentiums).
We have reported only preliminary results to the client as yet, so no conclusions here!