The Black Hole Illusion: Theoretical and Empirical Implications
thesisposted on 2021-05-07, 17:08 authored by Victoria L JakicicVictoria L Jakicic
The Black Hole Illusion (BHI) is a nighttime aviation landing illusion that occurs when only the runway is visible to pilots. With this illusion, pilots overestimate their descent angle, which causes them to overestimate their height, compensate by flying lower, and crash into the ground. Perrone (1983) proved that, in daylight conditions, it is possible to compute the descent angle using a ratio of retinal distances of the runway. He also showed that using a modified algorithm, which applies in nighttime conditions with just the visible runway, pilots would overestimate the descent angle, ultimately leading to the BHI. We quantitatively analyzed Perrone’s algorithm to determine whether it is a viable explanation for the BHI. Our analysis finds that, if Perrone’s algorithm was a complete explanation of the BHI, pilots would sometimes crash plane nearly half a kilometer before the start of the runway, which is much larger than what has been observed. To further investigate the role of Perrone’s algorithm, we derived and tested two additional predictions. First, we quantified an observation made by Perrone that the BHI illusion magnitude should be affected by the runway width. Second, we discovered that some conditions of runway width and length predict a reverse BHI such that pilots should underestimate their angle of descent and compensate by flying higher (and landing later) than is necessary. We empirically tested our predictions with an on-line experiment using simulated landing flight paths. Participants observed a cockpit view of a runway during five seconds of approach. In a subsequent still image, the participants indicated where they believed the plane would land if it continued its flight path. We measured the accuracy of landing positions for various runway widths and for various background contexts, specifically no-context, partial-context, and full-context backgrounds. Based on our analyses, predictions, and experiments, we cannot conclude that Perrone’s algorithm is what pilots use while landing an aircraft.