| Many years of research have been devoted to examining the limits of our visual system. Not only are there limits in the amount of information we can attend to, there also are limits in the number of items we retain from one moment to the next. Many experiments have used change detection paradigms to study the capacity limits of short-term memory, but none have looked at how relevant this limited capacity is in incidental change detection studies. In addition, previous research has not investigated short-term memory load and change detection in a dual task paradigm, in which change detection is not the primary task. Experiment 1 focused on global working memory load and incidental change detection. A general load effect was found such that observers who were asked to complete a verbal processing task detected fewer changes than observers who did not complete the verbal processing task. In Experiment 2, a specific load task was used as the primary task in conjunction with incidental change detection. In addition, a recognition lineup task was added to examine representations for the changing objects. Observers showed higher levels of change blindness when they searched for a more complex visual stimulus than when they searched for a simpler visual stimulus. In Experiment 3, observers searched for a visual stimulus that was either from the same dimension as the changing object (i.e. color) or from a different dimension than the changing object (i.e. shape). The type of visual stimulus did not affect change detection. In Experiment 4, methodological changes were made to replicate the findings from Experiment 2 and to further explore the relationship between incidental change detection and the cue search task. However, the complexity effect from Experiment 2 was not replicated. As such, Experiments 5 and 6 examined change detection with a more difficult concurrent visual memory task. There was no difference in change detection when observers searched for several cues compared to when they searched for just one cue. In general, a working memory load effect was demonstrated only when the concurrent task involved the storage and manipulation of information. |
