Abstract
Visual search performance improves when a linear boundary in feature space separates the target from the distractors. Xu et al. (2021) proposed an interpretation of this effect based on the observation that distractor heterogeneity is always higher in non-separable than in linearly-separable conditions. Using the Distractor Heterogeneity equation of Wang et al. (2017), Xu et al. (2021) found that the performance in heterogeneous search in orientation feature space was a function of target-distractor similarity and inter-item interactions. Distractor Heterogeneity was able to predict 95-98% of the variance in heterogeneous search without including any factor indexing linear separability per se. This result suggests that the slower search rates observed in non-separable conditions might be due to higher distractor heterogeneity affecting the strength of inter-item interactions. The current study extends these findings to color space. Experiment 1 measured observers’ performance in homogeneous search for a target color among one type of distractor color. The parameters observed in Experiment 1 were then used to predict search times in Experiments 2 (separable condition) and 3 (non-separable condition), where the target was presented in heterogeneous displays containing two distractor colors. Distractor Heterogeneity accounted for 86%-92% of the total variance in these data. These findings emphasize that the difference in difficulty between separable and non-separable conditions is not categorical as previously thought. Rather, the degree of distractor heterogeneity seems to explain the different search performance in these conditions.