The analysis of traffic accident data, that is historical data, is made in order to guide our actions directed at the reduction of future accidents. In general we may state that the purpose of the study of accidents is to create the possibility of forecasting future accidents. Our interest in the tragedies of the past comes out of our interest in avoiding future tragedies. As has been shown in multidisciplinary accident investigations, · each single accident constitutes a small universe of personal conditions, physical conditions and random events. However, accidents occur in such numbers that all of these apparently unique conditions occur over and over again. Experience shows that these circumstances are recurrent events. Since, "History repeats itself," it is possible ' to predict accidents statistically. The treatment of accidents requires data reduction from many small universes to fairly simple relationships. This process is necessary because it will reveal common factors. The common factors rather than individual differences between accidents are the keys to prediction. The circumstances leading to accidents--including road user behavior-depend on physical and social conditions: traffic flow, road user categories, road layout, roadway conditions, ,traffic codes, law enforcement, etc. These and other factors have been studied in detail, so it is possible, at least in principle, to set up a relationship between expected accidents and external conditions. (Author/publisher)
Abstract