Cue-Driven Changes in Detection Strategies Reflect Trade-Offs in Strategic Efficiency

We examined performance in a cued detection task when a peripheral cue was either 50% or 100% valid, tested in separate experiments. We combined a Posner cueing taskwith a double factorialmanipulation of stimulus salience. Unlike previous investigations in which participants responded to either target, we employed an AND decision task in which a target-present decision required there to be a target at both locations. When the cue was 50% valid, all the participants adopted parallel exhaustive processing to detect redundant targets with unlimited to limited capacity. When the cue was 100% valid, three participants, who performed this experiment first, adopted serial exhaustive processing. By contrast, the participants who first performed the 50% validityexperiment continued to adopt parallel exhaustive processing. Capacity generally declined below a lower bound, suggesting extremely limited capacity. Our conclusion is that the validityof the cue affected processing strategy but participants could increase the relative efficiency of the parallel processing with practice.