The time course of task switching: a speed--accuracy trade-off analysis.

On each trial of this study, participants either switched between or repeated two simple, two-choice tasks involving either letter or digit classifications. Speed-accuracy trade-off (SAT) curves were obtained using the response-signal method of eliciting speeded responses at various response time lags after the presentation of the stimulus for the second task. The key finding from separate analyses of the three SAT-curve parameters (intercept, rate, and asymptote) was that the location of the intercept of the SAT function (i.e., the point at which responding rises above chance) was shifted upward for both short and long response-stimulus intervals under task-switch conditions but only when the responses associated with each of the letter-digit stimulus components were incompatible.