Abstract
Rare targets are often missed during visual search tasks—a phenomenon known as the “low prevalence effect” (LPE). The significance of this effect is demonstrated in various professions (e.g., airport security screeners, radiologists, and sonar operators) in which accurate and efficient search performance is critical for detecting threats. While prior research has focused primarily on searching a static scene, the present work explored the LPE in a dynamic visual search task using random motion. Participants completed a visual search task through an online crowdsourcing platform, Amazon Mechanical Turk. The task consisted of a search display featuring 25 items where a target letter “T”, presented at either a high (50%) or low (10%) prevalence rate, appeared among non-target offset “L”s. The search items moved across the display at a constant speed, each item in a random direction. Search items only changed directions upon colliding with/bouncing away from other search items. When items reached an edge of the display, they would reappear on the opposite edge of the display continuing in the same direction. On each trial the items moved for a total period of 7 seconds before the items disappeared and a new trial array was presented. Consistent with the static visual search literature, a marked decrease (~74%) in target detection was found during low prevalence search compared to the high prevalence condition (p=.039). These results indicate that dynamic random motion is as susceptible to the LPE as static search. Accordingly, successful LPE mitigation strategies within a static search context may also be applicable to dynamic search with random motion.