Abstract
When looking for any object from a category (i.e. category search) the variability in appearance between exemplars in a target category influences search performance, with low variability categories leading to shorter reaction times (e.g. Hout et al., 2017; Nako, Smith, & Eimer, 2015). Searchers may not develop (or choose to utilize) a specific search template from highly variable target categories like BOOT compared to low variability categories (wherein exemplars do not vary much in appearance; e.g. BANANA). This study examined how searchers can use expectations from a particular scene to facilitate performance when a target category has too much variability for a searcher to develop a useful template. One group of participants provided our measure of category variability by searching for the category exemplars used in the primary experiment. Categories were thereafter divided into high, medium, and low variability groups, using the average RTs of exemplars within each category. In the primary experimental task, another group of participants received a basic level category word cue (e.g. "BOOT") and searched for a single exemplar from the categories used in the first task. Preceding their search, there was a prime image of a scene (e.g. a ski slope) that was either congruent to the target exemplar (e.g. a ski boot) or incongruent (e.g. a cowboy boot). We found an interaction between prime condition and category variability. When the target category had high variability, participants were faster to respond to targets in the congruent condition than in the incongruent condition. There was no difference between congruent and incongruent trials when the target categories had low variability. The results indicate that context (i.e. a scene) can help searchers develop useful expectations about the appearance of their target in category search, but context is not necessarily beneficial when the target category is homogenous in appearance.
Meeting abstract presented at VSS 2018