Douglas Addleman¹, Viola Störmer²; ¹Gonzaga University, ²Dartmouth College

During visual search, information about upcoming targets is highly beneficial: Knowing that your next target is likely red helps you find it much faster. Information about distractors is also helpful, but less so, even when compared to similarly informative target cues. How do these effects depend on how reliably cues predict upcoming features? We addressed this in four experiments (N=50 each) manipulating the format (target vs. distractor) and informativeness (75% vs. 92%) of cues to the color of upcoming search items. In all experiments, participants searched for a target C (left or right gap) among seven distractor Cs (top or bottom gap). Each trial contained eight items (four in each of two colors) and was preceded by a colored cue that predicted the upcoming target color (Exp. 1, 75% informativeness; Exp. 2, 92% informativeness) or non-target color (Exp. 3, 75% informativeness; Exp. 4, 92% informativeness). The specific cue color changed every 108 trials, with four different colors used across the experiment. Participants knew the precise informativeness and format of their cues. In all experiments, accurate cues led to faster responses than inaccurate ones (ps<.01). However, cue format and informativeness strongly influenced cueing benefits. When comparing validly cued trials to invalid ones, people were 236ms faster for 75% valid target cues and 385ms faster for 92% valid target cues, but only 48ms faster for 75% valid distractor cues and 162ms faster for 92% valid distractor cues. All across-experiment cueing benefits were significantly different (ps<.01). These results show that the greater effectiveness of target compared to distractor cueing holds at multiple levels of cue informativeness. Perhaps more importantly, they show that increasing informativeness of cues makes cueing effects stronger, even though participants have strong incentives to use cues to benefit performance even at moderate levels of informativeness.

Acknowledgements: This research was supported in part by NSF Grant BCS-1850738 to V.S..