Confidence in a visual search task depends on target visibility

In many visual search tasks, searchers decide about when to terminate search themselves. This allows them to search exhaustively and to report their decision only once they have reached a desired level of confidence. Here, we studied what determines searchers’ decision about the target’s presence or absence and their confidence in this decision when they cannot control the duration of their search. We combined a time-limited visual search task with a confidence forced-choice task. Each trial contained two subsequent search intervals of 800 ms, in which participants had to search for a Gabor target in a noise background. The Gabor could be present or absent in each interval independently. At the end of each trial, participants first had to report if the Gabor was present or absent, separately for both intervals and finally to choose for which interval they felt more confident. To manipulate the difficulty of the search, the radius of the noise background was varied between 3° and 7.5° in one of the two intervals. Since the Gabor target was small and required high-acuity foveal vision, the hit rate to successfully detect it depended on the minimum distance between gaze and the Gabor during the search. When gaze did not come close to the Gabor, the hit rate and the false alarm rate were of similar magnitude and were correlated across participants, suggesting that these hits were lucky guesses rather than true detections. Confidence choices on average preferred target-present over target-absent decisions. They also depended on the minimum distance between gaze and the Gabor and they were correlated with the hit rate within participants. These results indicate that participants’ decision about the target’s presence and their confidence in that decision are influenced similarly by how well the target was visible.