Can we predict perceived risk?

This paper examines the possibility that we may be able to use subjective measures of perceived risk in order to assess the relative safety of different shift systems. A large-scale survey of safety-critical engineers included three items relating to risk on each shift, namely alertness, likelihood of making a mistake and confidence in driving home after it. These three measures were found to load on a separate factor for each shift. Hierarchical regression analyses indicated that perceived night shift risk could be predicted on the basis of circadian type, the extent to which the engineers could control their work schedule, and a number of features of the scheduled shift system. In most cases the relationships observed were reasonably consistent with established trends in either performance capability or accident and injury frequency. However, there were exceptions to this indicating that results based on measures of perceived risk should be interpreted with the utmost caution.