Visual detection of line segments: Two exceptions to the object superiority effect

Boundary conditions for the object superiority effect (OSE, target lines embedded in meaningful drawings are more accurately identified than the same targets embedded in nonmeaningful drawings; Weisstein & Harris, 1974) were sought by examining two variables known to affect the word superiority effect. A robust OSE was obtained in both two- and four-alternative tasks with no mask or speed requirement. The OSE was eliminated by either a random masking stimulus composed of dots or lines (Experiment 1) or by requiring speeded identification responses (Experiment 2). An information processing model is proposed to account for the OSE, as well as the exceptions reported herein. It is shown that this model can be modified to deal with the analogous word superiority effect.