Global - Not Local - Variance Impacts Search
36.436, Sunday, May 12, 2:45 - 6:45 pm, Orchid Ballroom
Steve Haroz1, David Whitney2; 1University of California at Davis, 2University of California at Berkeley
When searching for an oddball target, how does the amount of visual variety (e.g. number of colors) immediately surrounding the target impact performance? We found that variance in the whole scene rather than just the local region matters. Subjects saw a display with 64 objects. On half of the trials, one object was given a unique visual property (e.g. one red target among many blue and green objects). The possible visual features that distinguished the objects included color, shape, or motion. Each feature had its own block of trials, so subjects knew which visual feature would contain the oddball. Subjects responded whether the target was present or absent (actually present 50%), and reaction time was recorded. Increasing the number of variants within the visual feature (e.g., going from red and blue to red, blue, green, and yellow) throughout the display (global variance) significantly impacted performance. However, increasing the number of variants immediately surrounding the target (local variance) while keeping the global variance constant showed no significant effect. Therefore, goal-directed or guided visual search appears to be susceptible to global display variance (how many distractor categories there are, and whether they are grouped) but is not sensitive to the local variance around the target itself.