Abstract
Abundant evidence in the attention literature has been interpreted to suggest that humans routinely use symbolic spatial cues like arrows or directional words to guide visual attention in everyday life. However, this literature has not considered the wide range of factors that influence voluntary choice behavior. Drawing from the rapidly growing decision making literature, we propose a rational analysis account of voluntary symbolic attention control – the Least Costs Hypothesis (LCH) – which construes voluntary control as a decision between intentional cue use and unguided search. In this view, observers deliberately pursue the strategy associated with fewer anticipated costs. The present study employed a spatial cuing paradigm and inferred voluntary cue use by the magnitude of "unbiased costs-plus-benefits," with larger magnitude effects reflecting greater cue usage. Consistent with the LCH, observers' decision was found to be sensitive to variations in cue processing efficiency. In Experiment 1, observers demonstrated a robust preference for using "easy-to-process" arrow cues to satisfy an easy visual search goal, relative to "hard-to-process" spatial word cues. Experiment 2 showed that this preference persisted even when the temporal costs of cue processing were neutralized (Figures 1 and 2). In Experiment 3, observers reported this cue type preference outside the context of a speeded task, and Experiment 4 showed empirical measures of this bias to be relatively stable over the duration of the experiment (Figure 3). Together with other evidence suggesting that observers' decision between intentional cue use and unguided search is also influenced by practice with the task and variations in unguided search efficiency, these findings suggest that voluntary symbolic attention control is mediated by ongoing metacognitive evaluations of demand that are sensitive to perceived variations in the time, effort, and opportunity costs of each strategy. Thus, voluntary symbolic attention control is far more complex than previously held.
Meeting abstract presented at VSS 2018