关键词: 视觉搜索, 靶概率效应, 优势反应, 抑制控制

Abstract: More recently, some researches showed that missing error rates were far higher at 1% target prevalence than at 50% prevalence (Wolfe et al., 2005; 2007; Fleck & Mitroff, 2007). Fleck and Mitroff (2007) suggested giving observers an opportunity to correct their last response can decrease the missing rates. However, the missing rates were still high before incorporating the correction responses. In other words, the mistake had been made before correcting it (Fleck and Midriff, 2007). Can the missing error be avoided in a more positive manner? The purpose of the present study was to find a better way to reduce missing errors in the visual search task.
We used a similar target detection task, in which the task of observers was to search tools (targets). In Experiment 1, 3, and 5, observers were required to press one of two keys as a response to the target or nontarget, while they were required to count or make markers in the Experiment 2 or 4 as a response to the targets. In Experiment 6, two tasks were implemented. One was a search task that was the same to Experiment 3, and the other was an inhibition control task.
The results of Experiment 1 and Experiment 3 indicated that the missing error was significantly increased when the prevalence of targets was low, which clearly replicated the differential prevalence effects on search performance identified in previous studies (Ethell & Manning, 2001; Egglin & Feinstein, 1996; Wolf et al., 2005). As observers in Experiment 1 reported, they noticed the targets in the display but they failed to shift response to the target pictures before pressing a key. Their claims were strongly supported by the results of Experiment 5, in which the observer’s eye movements were tracked.
Since that the response-execution is the main factor that caused the high missing error in low prevalence condition, the missing error was expected to decrease in the following three occasions. First, the missing error will be low when prepotent response is not induced in the target detection task. The results of Experiment 2 proved this expectation. Second, the missing error will be low when observers responded to the targets appropriately. This view is similar to the claims of Fleck and his colleagues (2007). The results of Experiment 4 in the present study suggested a more positive way to avoid the high missing errors. When observers applied an appropriated responding manner such as making markers, then the mistakes were avoided effectively. Third, the missing error will be low when observers possess a higher ability of execution control. This expectation was demonstrated in Experiment 6.
In sum, the results of the current study fully replicated the prevalence effect on visual search performance (Wolfe et al, 2005; 2007), which should alleviate concerns about methodological differences between studies. However, our results indicated that the high missing error can be avoided by applying the more appropriate responding pattern such as making markers and by recruiting the observers who have a higher ability of inhibition control.

Key words: visual search, prevalence effect, prepotent response, inhibition control