Abstract
The de-prioritization of old (previewed) items already in the field enables increased search efficiency for new items in the preview search paradigm (Watson & Humphreys, 1997). This effect has been replicated using valenced schematic faces (Blagrove & Watson, submitted). Previous work shows that changes to old previewed items that are likely to be of behavioural relevance (e.g., shape changes made when the new items are added), cause the old items to re-compete for selection with the new items. In contrast, less behaviourally relevant changes (e.g., color or luminance changes) have no effect and the old items remain suppressed. Here we examined the effect of changes in emotional valence to old previewed items. Specifically, we compared the detection of a schematic face target (either positively or negatively valenced) in a Preview condition (neutral preview, changing to a valenced distractor set) with a Half Element Baseline (neutral distractors only) and Full Element Baseline (mixed valenced and neutral distractor set). A neutral to negative change in the valence of previewed faces (Experiment 1) saw the preview benefit abolished, with search efficiency impaired, relative to both baseline conditions. In contrast, a neutral to positive preview change (Experiment 2) demonstrated a partial preview benefit, where participants were more efficient than in the Full Element Baseline. This contrast across experiments was also reflected in the numerical strength of the Preview Benefit, via an index of Preview search efficiency (PEI) (Experiment 1: PEI=0.08; Experiment 2: PEI= 0.41). Overall, the findings show that ignoring a negative change in expression is more difficult than ignoring a positive change in expression, consistent with ecological considerations.