Using Transportation Accident Databases in Human Factors Research

Accident databases commonly contain factual information about the time and date of each accident, vehicle characteristics, number of persons killed and injured, and other kinds of factual data. These attributes of the environment and equipment are usually directly represented in databases. In contrast, detailed analysis of accident causes, including human factors information, are frequently not represented because they are much more difficult to obtain and code.

This paper explores the suitability of transportation accident databases for use in human factors research. Given the goal of reducing the number and severity of transportation accidents, it is useful to use accident data as a tool to understand the common causes of accidents. Problems arise, however, because existing databases were typically not created explicitly for research purposes, and coding systems and file structures often omit or obscure useful information. Improved coding schemes and file structures that promote the use of databases for human factors research are discussed. Accident investigation methodologies that can improve the quality of human factors information in databases are also considered. Finally, problems associated with the use of existing databases are noted.