Abstract
Since an initial demonstration by Duncan (1984), many studies have found that it is easier to divide attention between two features belonging to the same object, than between two features of different objects—i.e., a “same-object benefit” (SOB). However, other studies, notably by Davis and colleagues (e.g., Davis & Holmes, 2005), have consistently found the opposite effect—i.e., a “different-object benefit” (DOB). We have replicated both effects repeatedly in our laboratory (e.g., Burnham & Neill, 2006), with the critical variable appearing to be the shape and/or complexity of the objects. Whereas most experiments finding SOB have used simple rectangles, experiments finding DOB have used more complex U-shaped objects. Our current experiments are designed to explore what object characteristics are necessary for either SOB or DOB to occur. Most SOB and DOB experiments have emphasized reaction time to clearly visible target stimuli. We have begun to test SOB or DOB for judgments of degraded (masked) stimuli, with Signal-Detection Theory (SDT) analyses of accuracy. In a recent experiment, we randomly intermixed trials with pairs of rectangles and trials with pairs of U-shapes. Subjects made unspeeded same/different judgments of two letters (E or F), briefly presented in ends of either the same object or different objects (equated for physical distance). Surprisingly, U-shapes yielded SOB in discriminability (d' = 1.54 vs. 1.40) virtually identical to rectangles (d' = 1.58 vs. 1.40). ANOVA yielded a highly significant main effect of same- versus different object, F(1,27) = 55.63, p .26. Because this is the first experiment in which we (a) randomized shapes over trials, and (b) measured accuracy of unspeeded responses, we are now testing accuracy when the shapes are blocked, and (conversely) reaction time when the shapes are randomized.