Carl Westin

6-2 Decision-making in ATC conflict resolution 119 6-2-1 Consistency in conflict resolution Although no literature was found on how consistent controllers are in conflict res- olution, a few studies addressing consistency in relation to other ATC tasks were found. In a large survey involving 100 controllers, almost a third of all respondents reported that they try to consistently use the same techniques and patterns when looking for information and scanning a sector. 189 In contrast, research by Redding et al. indicates that controllers vary their use of decision-making strategies during work, 40 which points to inconsistency. Controller consistency has also been investigated by means of assessing exper- tise using the Cochran-Weiss-Shanteau (CWS) index, 248, 249 derived by calculat- ing the ratio of discrimination (degree to which similar stimuli are discriminated from one another) to consistency (degree to which a repeated stimulus is judged the same). In a high-fidelity simulation involving twelve controllers, Thomas et al. 250 found a moderate correlation between the degree of controllers’ consistency, as measured by the CWS index, and simulator measures of their performance and efficiency. Controllers performed less consistently with deteriorating performance measured by the number and duration of separation errors. Similarly, they per- formed less consistently with decreasing task efficiency measured by the number of altitude and heading instructions issued. Although results indicated that consistency also decreased with increased complexity (measured by the number of aircraft con- trolled and handed off) the correlation was not significant. In addition, the authors noted that some controllers were considerably more consistent than others. In a study investigating the consistency in traffic complexity judgments, con- trollers were asked to pairwise compare seven static traffic scenarios and identify the most complex one in each pair. The method for determining consistency based on analyzing pairwise comparisons was derived from the Analytic Hierarchy Pro- cess (AHP). 251 Controllers were found inconsistent in their judgments with only one out of five found consistent according to the AHP’s suggested threshold of attain- ing a consistency ratio below 0.1 (with 0 indicating that judgments were perfectly consistent and 1 indicating that they were purely random). 252 6-2-2 Agreement in conflict resolution Research suggests that there may be considerable variability between controllers’ judgment and decision-making preferences in CD&R, 35, 131 especially between ATC facilities, 189 regions, and nations. 35 In the development of the conflict resolution system ISAC, differences in controllers’ view of the same conflict complicated the eliciting process of parameters underlying the system’s case-based reasoning model for conflict resolution. 131 Validation trails with other controller-based CD&R automation have shown that these systems do not match the variability in con-