Singleton search is guided by knowledge of the target, but maybe it shouldn't be

Among studies of visual search for a singleton, some studies show evidence of top-down attentional guidance driven by goals, while others do not, leading to uncertainty as to how goal-driven guidance should be included in attentional theories. Six experiments tested this guidance when a target shape is found by locating a singleton feature (color or shape) and an orientation within the target is then reported. Experiments 1 and 2 use the dimensional priming paradigm underlying the most effective arguments against goal-driven guidance, and show evidence for guidance in many circumstances. Experiment 3 extends the results to feature priming, and demonstrates a complex interaction between attentional goals and memory for previous targets. In Experiment 4, symbolic (word) cues were just as effective as image cues, further strengthening the case for goal-driven guidance. In Experiments 5 and 6, as in the previous experiments, valid cues again produced faster responses than invalid cues, showing the advantage of goal-driven guidance. Surprisingly, however, responses were even faster when the cues were uninformative. Furthermore, participants who began the experiment with neutral cues seemed to ignore informative cues later in the experiment. The results show that attention can be guided by goals even in easy searches, but that searchers have much flexibility in the use of this guidance, and may choose not to use it. Furthermore, their decisions about using this guidance are not always well informed, because they are not aware of the relative costs and benefits.