“…Also, when a two-dimensional, symbolic stimulus set (e.g., two letters, O and Z, in large and small sizes) is assigned to the four keypress responses, RTs are faster for mappings in which the more salient stimulus feature (in this case, letter identity, as established independently by faster identity discriminations than size discriminations in two-choice tasks) distinguishes the two leftmost and two rightmost response locations than for mappings in which it distinguishes alternating locations (e.g., Dutta & Proctor, 1992;Miller, 1982;Proctor & Reeve, 1985;Proctor, Reeve, Weeks, Dornier, & Van Zandt, 1991). Similar results are obtained when four stimulus locations are assigned to a two-dimensional, vocal response set (Proctor, Dutta, Kelly, & Weeks, 1994). Thus there is considerable evidence that for linear arrays of four stimuli and/or four responses, the distinction between the left-right locations is more salient than that between the alternate locations.…”