原文传递 Elucidating Visual Search Behavior and Resulting Performance in Low Prevalence Dynamic Search.
题名: Elucidating Visual Search Behavior and Resulting Performance in Low Prevalence Dynamic Search.
作者: Diaz, K; Wise, M; Guillory, S; Bolkhovsky, J; Peltier, C.
摘要: Visual search is an essential element of various professions, such as airport security screening, radiology, and sonar monitoring, in which accurate and efficient taskperformance is vital. Visual search performance is dependent on specific search conditions, such as how many targets are present, if targets are moving or stationary,and how long the targets are visible. Despite existing research on how conditions such as these individually influence search performance, the extent to which theyinteract—particularly target prevalence in different types of search environments—is unclear. Therefore, this study investigated performance and perceived cognitivedemands of three common aspects of visual search: target prevalence (50% vs 10%), display type (static still-image vs. dynamic continuous scroll), and display speed(3.5 vs. 7 seconds). Participants (n = 556) recruited through Amazon Mechanical Turk were randomly assigned to one of eight experimental conditions created fromthe combinations of these search conditions. Participants were tasked with detecting known targets, “T”s, among a display of distractors, offset “L”s, andsubsequently rating their perceived condition-specific cognitive task workload via the NASA Task Load Index (NASA-TLX). Though no significant interactionsbetween search conditions were found, condition-specific effects were present. Lower target prevalence produced decreased hit rates, replicating the Low PrevalenceEffect (LPE) (Wolfe et al., 2007), across static and dynamic displays. Additionally, NASA-TLX ratings implies observer blindness to actual performance decrementsin low target prevalence searches. The persistence of the LPE highlights the need for LPE-mitigation methods in both types of search environments.
总页数: 35 pages
检索历史
应用推荐