Speed-accuracy tradeoff during response preparation.

The speed-accuracy operating curve was investigated in a movement precuing two- or four-choice reaction time task. Four levels of response preferences were manipulated with subject instructions and postresponse information: (a) accuracy, (b) reaction time latency, (c) accuracy and reaction time latency, and (d) no preference. Eighty subjects completed 480 discrete keypressing responses with the index and middle fingers of both hands. The mixed design mean reaction time analysis indicated faster performances for the reaction time latency and the accuracy and reaction time latency groups than the no preference group. Additionally, the percent correct analysis revealed two significant interactions: (a) Trial Block x Precue x Response Preference, and (b) Delay x Precue x Hand Position. Overall, the present findings provide partial support for the speed-accuracy operating curve predictions. Caution is advised when drawing chronometric inferences based only on reaction time data or when response accuracies are extremely high.