Error models for accident studies in man-machine systems. Paper presented to the Fourth International Conference on the Foundation and Applications of Utility, Risk and Decision Theory, Budapest, June 6-10, 1988.

The analysis of accident causation is a theoretically and socially important problem-solving task. There is a general need for people to get a reasonable explanation of how and why socalled impossible and unforseen accidents did occur. Only with this kind of knowledge people are motivated and able to think about preventive measures. This paper focuses upon different error models and discusses the usefulness of error analysis for a better understanding of the human factor in accident causation. Arguments are presented to indicate that evaluations of risks play only a marginal role in determining routine behaviour. Wrong hypotheses and habits are the main sources of error, implicating that in error prevention lack of understanding rather than lack of motivation is the problem to be dealt with.