| Summary: | 碩士 === 臺灣大學 === 心理學研究所 === 98 === The objective of this study is to understand whether recollection and familiarity processes are both involved in recognizing a pre-change object under change blindness and change detected conditions, and whether they are affected by the completeness of representation. We manipulated viewing duration between two experiments (180 ms in Experiment 1 and 2 s in Experiment 2) to vary the completeness of representation. In both Experiments 1A and 2A, participants were required to detect a change between two arrays and then performed either an inclusion or exclusion recognition task. We used the modified dual-process signal detection model and a Bayesian method to estimate the proportion of recollection and familiarity under change detected and change blindness conditions. The results from both experiments supported the involvement of dual processes in recognizing a pre-change object. The proportion of recollection process was higher under correct detection than under change blindness and was higher when representations were more complete. To further verify the information recollected prior to performing change detection, we conducted Experiments 1B and 2B in which participants performed recognition prior to carrying out the detection task. Participants were also asked to report the contextual information upon which they performed the recognition task. When viewing duration was brief in Experiment 1B, temporal order and object category were the primary sources recollected under correct detection. When viewing duration was long in Experiment 2B, participants recollected temporal, object category, spatial location, and perceptual details under correct detection. These results suggested that successful comparison with a post-change object can be made when more sources of a pre-change object were recollected. Recollection failure is a reason for change blindness.
|
|---|
