Response to an Automated Function Failure Cue: An Operational Measure of Complacency

Abstract

This report describes an experiment whose purpose was to operationally define complacency. Non-pilots performed an airspeed and altitude maintenance task on a desktop approach procedures trainer, while responding to three unpredictable reaction time cues each minute during a 20 minute trial. A separate reaction time cue was yoked to come on with this cue and extinguish automatically with it; but subjects were instructed that they had to respond to this second cue if it failed to extinguish by itself. This otherwise irrelevant cue simulated an automated function whose persistence signaled a system failure and called for detection and response on the part of the subject. The delay in response to this “complacency” probe, compared with the latencies of the ongoing reaction time task, comprised the measure of complacency. Complacency reaction time was significantly longer and, while a second complacency probe led to improved response times, evidence for complacency persisted. Non-pilot subjects tended to keep the trainer within 100 feet of the assigned altitude, and within 10 knots of the assigned airspeed. There was no evidence of degradation of performance of either the flying tasks or the reaction times as a function of time-sharing.