Results

Plan-View Display

The strategy that most of the test participants used to determine the descending aircraft was a systematic and exhaustive search across the display screen. This strategy ensured complete coverage and minimized the possibility that any candidates had been missed. Since climb rates remained steady for the duration of the scenario and the "climb rate" readout in the hook window was unambiguous, there was no need to return to a track if it had been correctly checked before. Therefore, this was an effective strategy and allowed most participants to search the entire display only once in order to complete the task correctly. Other less efficient strategies were employed. Two participants searched in quadrants, sometimes reexamining symbols on the quadrant borders. Two participants searched all unclustered symbols first and then returned to the clusters. One participant searched the symbols according to ID (shape), unknowns first and then friends.

The tables of results for Scenarios A and B show the number of correctly marked descending aircraft for each of the 14 participants as a function of time. The second row of both of these tables shows the total time to completion for each of the 14 participants with the mean and standard deviation at the end of the row.

As the data plots of the mean cumulative number of descending aircraft correctly selected versus time for Scenarios A and B show, the search technique that the participants used had a considerable impact on the rate at which the descending aircraft were found and marked. The nearly constant slope of the curves up to about the 80 percent point implies that the participants discovered the descenders at a rapid and nearly constant rate. After that point, the slopes flatten out, signaling a decrease in the discovery rate. This is a phenomenon that can also be explained by the search technique. If the participants found that they had missed one or more of the descenders, they would be forced to re-search the entire display to find the one or two they had originally missed.

The need to re-search the entire display in some cases, along with the variations in the inherent abilities of the participants to master the task, produced a wide variance in the times to completion for the 14 participants. The mean times to fully complete the task were 129.8 seconds with a standard deviation of 60.8 for Scenario A and 172 seconds with a standard deviation of 92 for Scenario B. The mean times to reach 70% completion of the task were 82.1 seconds with a standard deviation of 20.9 for Scenario A and 80.4 seconds with a standard deviation of 19.2 for Scenario B.

Another notable feature of these results is the almost 50 percent increase in the average time to full completion from Scenario A (10 descenders) to Scenario B (21 descenders). The test administrator observed that Scenario B had several clusters of aircraft which caused the participants to be much more careful, and therefore slower, in hooking the aircraft symbols in these clusters. Also, the 21 descenders in Scenario B presented a high probability that one or more would be missed and, therefore, a higher probability that they would need to re-search the display to find the missed descender.

Perspective Display

Most participants used a somewhat less structured method to search the perspective display for descending aircraft. The typical method was to quickly mark those aircraft that were obviously descending, and then to more slowly and carefully examine the remaining tracks for pitch attitude cues. Most participants initially attempted a more structured scan, but were easily diverted by obvious descenders. Participants varied in their confidence of the pitch attitude cues. Most participants would tag obvious descenders without hesitation. Two participants did not tag any descenders without first double checking with the hook panel. Responses to steady aircraft and harder to judge descenders ranged from a more lengthy visual examination to double checking with the hook readout. Ascenders were rarely hooked. Only two participants did not utilize the hook readout at all.

The tables of results for Scenarios A and B show the number of correctly marked descending aircraft for each of the 14 participants as a function of time. As before, the second row of both of these tables shows the total time to completion for each of the 14 participants with the mean and standard deviation at the end of the row.

Several of the participants experienced considerable difficulty in finding the last 2 or 3 descenders, and this fact is reflected in the plots of the mean time to completion data of Scenarios A and B, respectively. These plots show a steep and fairly constant rise in the number of descenders correctly marked, until the total reaches about 70 to 80 percent of the available descenders. At this point the plots flatten out, indicating that the participants were having difficulty, and were, therefore, taking more time, in finding the last 20 to 30 percent of the descenders.

The mean times to fully complete the task for the perspective display were 198.2 seconds for Scenario A and 228.4 seconds for Scenario B, with standard deviations of 122.8 and 102.8 seconds, respectively. The mean times to 70% completion of the task for the perspective display were 47.5 seconds for Scenario A and 78.2 seconds for Scenario B, with standard deviations of 19.2 and 23.1 seconds, respectively.

Performance Comparison Between Displays

Analyses of variance (ANOVAs) were performed on both the time-to-completion data and the time-to-70-percent data, to determine if there were any significant differences in performance between the plan-view and perspective displays, and between the two scenarios. For each ANOVA, there were 56 data points; one for each combination of Scenario and display type for each of the 14 experimental participants.

The time-to-completion analysis shows that display type has a significant effect on mean time to completion: F = 5.76, p = .02. The Scenario had no significant effect: F = 1.94, p = .17. The plan-view display performance was better than the perspective display performance.

The time-to-70-percent analysis shows that display type again had a significant effect on operator performance: F = 11.11, p = .002. Note that in this case, the perspective display was significantly better than the plan-view display. Here the Scenario had a significant effect: F = 6.87, p = .01.

The time to 70 percent can be superimposed over the time to completion to get an indication of the difficulty each participant had in finding the last 30 percent of the descenders. For the plan-view display, the graphs for Scenario A and the graphs for Scenario B show that most participants exhibited nearly linear performance, having no more difficulty in finding the last 30 percent than the first 70 percent. Spikes in the 100 percent completion area which are two or three times the height of the corresponding part of the 70 percent region indicate where participants missed a descender and had to re-search the entire scenario. For the perspective display, the graphs for Scenario A and the graphs for Scenario B show that most participants had extreme difficulty in finding the last 30 percent of the descenders. Spikes in the 100 percent completion area indicate where participants missed a descender and had to ponder the visual cues or resort to hooking.

Continue

Table of Contents