Document Type


Publication Date





The relationship between communication events and controller workload has been well established. Unfortunately, a substantial amount of time and effort is required to transcribe and code these events. Alternative measures might be preferable if they could be obtained more easily. Manning, Mills, Fox, Pfleiderer, and Mogilka (2002) found that, relative to a set of computer-derived measures, communication events might not add enough unique information to the prediction of subjective workload to justify the effort involved in obtaining them. At the time the study was conducted, computer-derived measures of altitude, heading, and speed changes were not available. The present investigation compares altitude, heading, and speed clearances with computer-derived measures of altitude, heading, and speed changes. Two 20-minute samples of live air traffic data were collected from each of four sectors in the Kansas City en route airspace. Communications data were transcribed from audio recordings and coded (e.g., altitude, heading, and speed clearances). Altitude, heading, and speed changes for each of the 20-minute samples were computed using the Performance and Objective Workload Evaluation Research (POWER) software system. The 20-minute samples were parsed into 4-minute intervals, and the number of communications events and changes were tallied for each interval. In addition, 16 subject-matter experts provided Air Traffic Workload Input Technique (ATWIT) measures for each 4-minute interval for all samples. Multiple regression analysis of altitude, heading, and speed clearances on mean ATWIT scores yielded an R = .59 (R2 = .35). Multiple regression of the number of computer-detected altitude, heading, and speed changes on mean ATWIT scores yielded the same results. Multiple regression of both clearances and changes, employed to examine shared and unique variance of the two sets of measures, revealed that altitude changes alone could account for most of the variance in ATWIT scores (R = .67; R2 = .44). Results suggest that computer-derived measures of altitude and heading changes may be a viable substitute for more labor-intensive communication measures. However, ground speeds recorded by the Host computer (and displayed on the controllers’ radarscope) were too erratic to provide a valid measure of speed changes and could not be recommended as an acceptable alternative for speed clearances.