Relationships of performance accuracy, target call sign sensitivity, and response bias on the Coordinated Response Measure task

Abstract

Coordinated Response Measure tasks evaluate attentional processing by requiring the listener to mark the color-number response contained in a phrase that starts with a specific call sign. However, when a target call sign is present on every trial, it is not known whether subjects must perceive the call sign in order to respond correctly. The present experiment examined the relationships among sensitivity to the target call sign, response bias, and performance accuracy in 6 normal-hearing subjects over 4 testing sessions. The number of talkers (2–4) and talker sex configuration were manipulated in 9 stimulus conditions. In session 1, a target call sign was present on every trial. During sessions 2 and 4, 10% of the trials were foils with no target call sign, while the stimuli in the third session had an equal number of trials with and without the target call sign. Results indicate significant main effects of test session and sex configuration for all 3 experimental parameters. High correlations were observed between sensitivity and accuracy for all sessions. However, a significant improvement in sensitivity was not accompanied by a similar improvement in accuracy, suggesting that better call sign detection does not necessarily lead to better task performance.