Abstract
A conjunction search target can be distinguished from distractors based on the combination of relevant features (e.g., a red horizontal bar, presented among green horizontal bars and red vertical bars). Olds and Fockler (2004) developed a modified conjunction search paradigm, where one of two relevant features, for each search item, was previewed immediately before the intact search display. In a colour-orientation-preview condition, the colour of each display item was previewed for 1 second, followed by a 1-second preview of the orientation of each display item. An orientation-colour-preview condition presented the orientation-preview display first. Observers' search of the immediately subsequent intact display was better if the colour-preview came before the orientation-preview than if the orientation-preview came first.
Related to this asymmetry, Hannus et al. (2006) matched feature discriminability for colour feature search, with feature discriminability for orientation feature search, and used those feature values to create a colour/orientation conjunction search. Observers made more saccades to items that matched the target colour than to items that matched the target orientation, even though those feature differences were matched for discriminability in feature search.
The present experiments replicated the results of Olds and Fockler (2004) with different coloured stimuli. Furthermore, a mixed-preview condition simultaneously previewed the colours of half of the display items, and the orientations of the other items; a mixed-mixed-preview condition presented an initial mixed-preview display, followed by another mixed-preview display in which the other feature of each item was previewed for 1 second. The mixed-mixed-preview failed to facilitate search, even though it included the same information as the colour-orientation-preview (1 second per feature, per item) which greatly facilitated search. The striking difference in facilitation, between previewing one feature at a time and previewing mixtures of features, is discussed in the context of the literature on perceptual grouping and scene segmentation.