The Effect of Collateral Alarms on Primary Response Behavior

Abstract

The growing complexity of aircraft systems has increased the likelihood for false alarms as well as multiple alarm occurrences. Understanding patterns of diagnostic and response behaviors to these alarms is important for system efficiency and safety. The present study was designed to examine whether inexperienced operators will utilize collateral alarms as a confirmation about the validity a given alarm, while ignoring the base rate probability for that alarm being true. Two experiments were conducted to determine whether participants’ confidence levels in a 50% true alarm would vary as a function of the number of collateral alarms. The procedures were similar for both experiments, in that zero to five collateral alarms were presented to participants along with a given 50% true alarm. However, while the first experiment was a repeated-measures design, the second experiment was conducted as a between-subjects design to insure that results of the first experiment were not an artifact of design. Both experiments yielded similar results, showing that inexperienced operators, when reporting their confidence in the validity of a given alarm, are influenced by the presence of other alarms. Moreover, overconfidence occurred when several collateral alarms were present, whereas under-confidence occurred when a minimum number were present. These findings indicate that collateral alarms may be used as a confirmation for alarm diagnostics by inexperienced operators, thereby they are assuming that multiple alarms are systematically related. Practical implications for training and effective alarm system design are discussed.