Searching for the alerting effect: the optimal SOA is longer in compound – than in simple – search tasks.

Simple visual search involves the single step of finding a target (e.g., a red ring) among a set of distractors (e.g., green rings). In contrast, compound search involves two steps. For example: i) find the target ring in the display, and ii) identify the orientation of a line segment inside the target. Performance is known to be facilitated when the search display is preceded by an alerting stimulus, such as a brief brightening of the screen. Until recently, alerting was studied using only “simple” tasks. In these studies, the optimal stimulus-onset asynchrony (SOA) between the alerting stimulus and the search display was found to be about 100 ms. Recent work that employed a 100-ms SOA showed that while alerting does occur in simple search, it does not occur in compound search. A temporal-period model was proposed to account for these findings. In the present work, we varied the SOA to test predictions from that model. In Experiment 1, we used a compound task with two SOAs: 100 and 150 ms. The results confirmed the absence of alerting when the SOA was 100 ms and revealed significant alerting when the SOA was 150 ms. To examine the time course of the effect, Experiment 2 included four SOAs: 50, 100, 150, and 200 ms. An alerting effect was found when the SOA was 150 ms, as in Experiment 1, but not when it was 100 or 200 ms. When the SOA was 50 ms, the alerting stimulus led to worse performance than when the alerting stimulus was absent. The temporal-period model was revised to account for this pattern of results.