Differentiating types of set-level compatibility

Two types of stimulus-response compatibility effects can be measured. Element-level effects are differences in reaction time (RT) for alternative mappings of the members of a stimulus set to the members of a response set. Set-level effects are differences in RT between various combinations of stimulus sets and response sets when the most compatible element-level mappings for each combination are compared. Kornblum, Hasbroucq, and Osman (1990) equated set-level compatibility with dimensional overlap and proposed a dimensional overlap model in which automatic activation of the response corresponding to a stimulus is an increasing function of dimensional overlap. This model thus predicts a positive relation between set- and element-level compatibilities. Although Kornblum et al. treat set-level compatibility as a unitary construct, there are conceptual distinctions that can be made between different types of set-level compatibility. We present three lines of research that use different ways of manipulating set-level compatibility. Each shows a distinct pattern of effects for the relation between set-and element-level compatibility, suggesting that different types of set-level manipulations may in turn affect different processes and not have the same influence on element-level compatibility effects. We discuss issues in measuring the relation between set- and element-level compatibility effects and propose that it is necessary to distinguish different types of set-level compatibility rather than combining them under the unitary construct of dimensional overlap. © 1997 Elsevier B.V. All rights reserved.