Abstract
The contingent spatial blink paradigm developed by Folk, Leber, and Egeth (2002) reveals a surprising lack of precision when it comes to our ability to maintain attention goals (sets). A completely irrelevant distractor sharing the color of a target letter in an RSVP (rapid serial visual presentation) stream produces a substantial attention capture effect, even when this distractor is distant from the RSVP stream, and the target's spatial location is central and constant. We examined whether the precision of attention sets can be improved by manipulating participants' expectations. In one experiment, the probability of the target-color-matching distractor was manipulated between groups to appear frequently or infrequently. We hypothesized that participants who encountered the peripheral distractor more frequently would develop a more precise attention set encompassing only the target's location, resulting in less capture. Consistent with this hypothesis, capture effects were smaller when the distractor's occurrence was frequent. We also manipulated the probability of the distractor occurring at a specific location, but found that consistent distractor location did not elicit less capture compared to when the distractor had a chance of occurring at four possible locations. Finally, we examined the possibility of training participants to adopt a temporally narrow attention set. Training involved searching for a target that occurred in the RSVP stream randomly (broad temporal window), or searching for a target that always occurred in the middle of the RSVP sequence (narrow temporal window). However, when participants were transferred to a block of trials in which the target always occurred in the middle of the stream and distractors could appear before or after it, capture effects were equivalent across groups. The present study suggests that experience (i.e., distractor frequency) can modify attention sets, but the precision of attention sets is largely unaffected by temporal and spatial expectations.