Abstract
The flanker paradigm has been used to study selective attention across many domains, and alternative theories have been proposed to account for the flanker effects observed across those different domains (e.g., selective perception, response competition, selective decision, perceptual crowding). To facilitate testing alternative theories, we describe a theoretical framework that is defined by multiple properties. Here we focus on two: Does a given flanker effect derive from processes within immediate perception, later processes or both? Does a given flanker effect arise from selection error, non-selective interactive processing or both? We investigated these properties within a single domain: spatial selection with a simple feature judgment. A target and flanker were presented in the periphery, distinguished by only a spatial cue. Observers discriminated the color of the target and ignored the color of the flanker. Accuracy was measured as a function of whether the target and flanker were response congruent or incongruent, and as a function of their separation. Results revealed a large congruency effect that depended on separation. To test whether this flanker effect derives from processes within immediate perception or later processes, we compared simultaneous and sequential displays of the target and flanker. There was little or no advantage for the sequential display, consistent with none of the effect deriving from immediate perception, but occurring within later processes. To test whether this flanker effect derives from errors in selection or non-selective interactive processing, we manipulated the quality of cue. The observed effects were accounted for by independent processing of the target and flanker, combined with selection errors. No need for interactive processing. For this flanker effect, we find evidence of selection errors (not non-selective interactive processing) in later processing (not immediate perception). We look forward to applying this analysis to other applications of the flanker paradigm.
Meeting abstract presented at VSS 2017