Problems in identifying causal factors

2 Statistical methods use accident statistics in the identification of causal factors. They may show that the probability of an accident will increase, if certain conditions are prevailing. These methods will be further discussed in Part III. 

We are here concerned with the identification of causal factors in individual accidents and near accidents for use in accident prevention. The analytic 

Checklists are applied in accident investigations in order to ensure that all relevant causal factors are considered. In practice, the checklist-supported identification of causal factors is problematic. Items at the top of the checklist are more often addressed than items further down (Hale et ai, 1997). Experience also shows that the identification of causal factors is affected more by the items on the checklists and the underlying accident model than by factual circumstances. A detailed presentation of certain causal aspects will result in an overestimation of the importance of such aspects. 

Supervisors have a tendency to choose causal factor alternatives that are not possible to verify and that involve limited management responsibilities and obligations to implement remedial actions (Kjellen, 1993). Typically, causal-factor alternatives related to human factors such as improper motivation are selected. The link between the identified causes and the selection of remedial actions is often weak. 

Example 2 In the same study, ten safety delegates were independently asked to code a written description of an accident by means of the ILCI model. No two such causal classifications were similar. 

---