The decision to eject: a receiver operating characteristic analysis

Although the technology of ejection-seat systems is highly developed and automated, the decision about whether to eject still rests with the pilot. One factor thought to affect that decision is the height at which the decision must be made. For this reason the Royal New Zealand Air Force considered lowering its designated ejection decision height of 10,000 ft above ground level so that pilots had more time to make a correct decision.

Lowering the ejection decision height will increase the accuracy of ejection decisions.

Thirty male Strike pilots each flew twenty simulated sorties. Half the sorties involved scenarios that required ejection and half did not. The pilots were divided into three equal groups which were randomly assigned to an ejection decision height of 10,000, 6,000, or 3,000 ft above ground level. After a sortie ended or was terminated by ejection, each pilot rated his confidence on a six-point scale that the scenario did or did not require ejection.

Receiver operating characteristic analysis showed that the ejection decision height did not influence the accuracy of the decision to eject, but did affect the bias for ejecting. The bias in favor of ejecting increased as the height was lowered. Experienced pilots were more decisive in making their judgments than less experienced pilots.

Lowering the ejection decision height did not affect the accuracy with which pilots could discriminate between flight scenarios that required ejection and ones that did not, but it did increase their bias to eject.