Quantitative projections of a quality measure: Performance of a complex task

Complex data series that arise during interaction between humans (operators) and advanced technology in a controlled and realistic setting have been explored. The purpose is to obtain quantitative measures that reflect quality in task performance: on a ship simulator, nine crews have solved the same exercise, and detailed maneuvering histories have been logged. There are many degrees of freedom, some of them connected to the fact that the vessels may be freely moved in any direction. To compare maneuvering histories, several measures were used: the time needed to reach the position of operation, the integrated angle between the hull direction and the direction of motion, and the extent of movement when the vessel is to be manually kept in a fixed position. These measures are expected to reflect quality in performance. We have also obtained expert quality evaluations of the crews. The quantitative measures and the expert evaluations, taken together, allow a ranking of crew performance. However, except for time and integrated angle, there is no correlation between the individual measures. This may indicate that complex situations with social and man–machine interactions need complex measures of quality in task performance. In general terms, we have established a context-dependent and flexible framework with quantitative measures in contact with a social-science concept that is hard to define. This approach may be useful for other (qualitative) concepts in social science that contain important information on the society.