How expectations alter search performance

Abstract

We assessed how expected search difficulty impacts search performance when expectations match and do not match reality. Expectations were manipulated using a blocked design (75% of trials presented at the expected difficulty; target–distractor similarity increased with difficulty). Expectancy was assessed by examining the change in search performance between trials with accurate expectations and easier-than-expected or harder-than-expected trials, matched for search difficulty. Observers searched for Landolt-C targets (Exp-1) or real-world objects (Exp-2). Increased difficulty resulted in reduced accuracy, increased RT and object dwell times (targets and distractors; both experiments), and reduced guidance (Exp-2). Relative to the same level of search difficulty and when expectations were accurate, harder-than-expected search reduced accuracy, RT, and target object dwell times (Exp-1). Whereas easier-than-expected search increased RT and target dwell times (Exp-1). While Experiment 2 showed somewhat muted expectancy effects, easier-than-expected search replicated the increased RT observed in Exp-1, with an additional guidance decrement and increased distractor dwell time. These results demonstrate that expectations shift search performance toward the expected difficulty level. Additionally, post hoc analyses revealed that observers who experience larger difficulty effects also experience larger expectancy effects in RT, guidance, and target dwell time.