Abstract
Extant accident classification systems have provided sound information for some phases of accident prevention research but no taxonomy has proved adequate for all areas of accident causation research. It is unlikely that a single classification system can meet all the needs of causation research. However, the authors believe the following factors-among others yet to be identified-are pertinent to causation, per se, responsibility; activity-situations in which accidents occur; and social nature of accidents. Analysis of the data suggests that such information is meaningful for accident causation research, and that further research along these lines should be purposed. A multidisciplinary is indicated.