The main aims of the research were to estimate the accident rate per year of exposure of stolen cars and to compare it with the known rate for the complete car fleet. This was achieved by using the Vehicle Registration Mark to link stolen car records, DVLA vehicle information and records of injury accidents from the STATS19 road accident database. The stolen car records from the Police National Computer suffer from one major limitation for this type of research - they do not include cars stolen and recovered on the same day, overnight or during a weekend. In the course of the analysis, it was found that this data set gave misleading results. To overcome this, complete stolen car records provided by four police forces in 1997 were analysed. Results showed that stolen cars were around four times more likely to be involved in an injury accident than a legitimately driven car. (Author/publisher)
Abstract