Abstract
In the real world, observers may search for more than one target at the same time. For example, baggage screeners search simultaneously for very different targets (guns, knives and IED). When observers search for two targets, some studies have found that performance is worse than if observers searched for only one. Moreover, Wolfe et al. (2005) showed that error rates are strongly influenced by prevalence. Does the relative prevalence of two types of target influence observers' accuracy and quitting rules? When is it time to stop searching if there are two things to look for? Our observers performed a visual search for an oriented target. Two types of targets were intermixed in a block. On each trial, there was one target or no target. Overall target prevalence was consistently 50% within a block. One target was the “standard” type. The other type could be easy, similar, or hard relative to the standard. The relative prevalence of the two targets could be 80%/20%, 50%/50%, or 20%/80% of target present trials in a block. Results: Prevalence had its largest effect on detection of the harder target. Target absent RTs were more strongly influenced by the target pair than by their relative prevalence. Standard-Easy RTs were shorter than Standard-Hard. What principle did observers use to determine search termination time? We developed a model, based on the assumption that observers are trying to maximize the number of targets found in a fixed period of time. Observers have some knowledge of their target-present RTs and of target prevalence. Based on this information, they set a termination time. The model is quite successful in capturing the pattern of the target-absent RTs.