Abolition of Asymmetry: the Impact of Pre-task on Visual Search
56.548, Tuesday, May 14, 2:45 - 6:45 pm, Vista Ballroom
Emily S. Cramer1, Ronald A. Rensink1; 1University of British Columbia
Visual search has often been assumed to be unaffected by tasks prior to search. However, we find that certain pre-tasks can abolish certain visual search asymmetries, making search equally inefficient for both types of target. To investigate this "abolition of Asymmetry" (AOA), two conditions were compared. In the first, 11 participants searched for a target defined by length. In the second, search was the same but was preceded by an object recognition (OR) task (Chua, Boland and Nisbett, 2005), which presented participants with a series of images and then probed their recollection of the images features. Observers not given a pre-task showed a strong asymmetry: 34.9ms/item for long targets vs 51.3 ms/item for short ones. However, when administered the pre-task, search rates became symmetric: 47.5 ms/item for long targets vs 50.3 ms/item for short ones. A possible explanation for this abolition is that the OR task induces a holistic style of visual processing, similar to that found in East Asian observers, who also show no asymmetries in search (Ueda et al., in preparation). To test this explanation, we conducted four further conditions. In the first we decreased the motivation for holistic processing by decreasing the size of the images in the OR Task. Symmetry was still abolished. In the second, we used the original pre-task, but imposed a frame over the search display. Both circular and square frames reinstated asymmetry, with the circle reinstating a slight (52.1ms/item to 40.8ms/item) and the square a strong (59.2ms/item to 42.1ms/item) asymmetry. In the last conditions, we examined two other types of visual search, replacing search for length with search for contrast and for a Q among O-shaped items. Both asymmetries survived the pre-task: as such, the pre-task appears to abolish asymmetry for only certain kinds of feature.