General agreement exists that the ultimate goals of traffic accident research are to reduce fatality, mitigate injury and decrease economic loss to society. Although massive quantities of data have been collected in local, national and international programs, attempts by analysts to use these data to explore ideas or support hypotheses have been met by a variety of problems. Specifically, the coded variables in the different files are not consistent and little information on accident etiology is collected. Examples of the inadequacies of present data in terms of the collected and coded variables are shown.
The vehicular, environmental and human (consisting of human factors and injury factors) variables are disproportionately represented in most existing data files in terms of recognized statistical evidence of accident causation. A systems approach is needed to identify critical, currently neglected variables and develop units of measurement and data collection procedures. The systems view would be based on the paradigm of the human operator in a man-machine system, simultaneously performing tracking and vigilance tasks and making decisions in a varying environment.
By increasing utilization of more sophisticated mathematical models of man-machine systems and controlled laboratory experiments, sets of critical variables could be defined and their interactions studied, both theoretically and experimentally. From these studies variables could be defined and hypotheses established for future field testing in traffic accident data collection and analysis programs, thereby more efficiently and effectively satisfying the original ultimate goals of traffic accident research.