The authors consider the problems of studying accident causation and put forward a procedure that they consider to be more effective. A combination of data collection and analysis procedures are described and the method illustrated by recent projects examining right-turn junction accidents and overtaking accidents. Data acquisition involves the use of police accident files from which information is extracted by coders. Two reliability studies are also described. The analysis process, aggregation and presentation are discussed. Some of the findings from the two projects are outlined. The use of these findings to identify strategies to prevent accidents is considered.
Abstract