Visual Search Improves with Detection Searches, Declines with Nondetection Search

摘要

Experiments examined improvements in target detection performance in visual search as a function of number of detection searches versus nondetection searches. The hypothesis questioned whether the number of times a detection state occurred versus the number of times the search process was executed determined the improvement rate. Subjects trained with consistently mapped (CM) targets with different numbers of detection opportunity searches versus nondetection opportunity (no target) searches. A multiple frame procedure was used in which subjects searched through 12 frames with 4 characters each for a single target. After training blocks subjects were tested with a target presented on every trial. The results show: (1) detection accuracy is primarily a function of number of detections as opposed to number of searches; (2) nondetection searches decrease detection accuracy; (3) as few as 10 CM detections can result in significant improvement in performance; (4) there may be some minimum processing time necessary after detection in order for the detection to result in improving later detection performance; and (5) the improvement from high detection probability training shows positive transfer to both low and high probability target search. (Author)