Research on errors and learning from them: Methodological perspectives

Abstract

Following a human error, one should not only know that something went wrong but it must also lead to the question of the underlying mechanism of the occurrence. Furthermore, learning from errors is only possible if the causes can be recognised. However, it is anything but easy to identify these mechanisms – whether a technical or a human error. Remarkably, in the scientific literature concerning human error, there are innumerable discussions about a generally accepted definition of human error: if errors are caused by humans at all, if they always have a cause, if systems can be designed to absorb human errors safely, etc. However, in these discussions, the question of an appropriate methodical approach seems to be a neglected concern. For this very reason, a selection of already known and used methodical approaches in research will be presented, especially historical experimental attempts. This chapter therefore reflects the question why the progress in error research is slow and why this topic is so complex. We focus on appropriate research methods for acting on errors, especially habitual errors. Methodical requirements are discussed on the basis of former analyses of this issue – Freud’s examinations and an exemplary contemporary consideration. The chapter concludes by first recommending training simulators as an approach to overcome analytical–methodical barriers, which are also favored in the authors’ own work. Second, an answer is given to the question of what can be learnt from errors.