Fig. 5.

Distribution of response errors and deviations from the locations of non-target items for different set sizes. (a) Frequency density of response errors (Euclidean distance from target location). (b) Frequency densities of deviations from non-target locations. Dots show mean frequency density for different error magnitudes, error bars show standard error over subjects. Dashed lines show the expected frequency densities under the assumption that non-target locations do not influence the response position; asterisks indicate significant deviations of empirical frequency from expected values (*p < .05; **p < .01). Simulation results for the best-fitting mixture model are plotted as solid lines.