S-R compatibility effects: Do we need a new theory?

Abstract

It has been known for many years now that reaction time (RT) to a stimulus is shorter when its relative spatial location corresponds to the relative location of the required response (e.g., Fitts and Seeger, 1953). This result, termed the stimulus-response (S-R) compatibility effect, has been explained by Simon (Simon, 1968; Simon & Rudell, 1967)as arising from a natural tendency to respond toward a source of stimulation. Compatible S-R pairs benefit from this tendency, while for incompatible pairs the tendency must be suppressed before correct responding can occur. Cotton, Tzeng, and Hardyck (1977, 1980) have recently reported a series of studies challenging this conceptualization. The task employed in a majority of these studies involved the subjects' responding to one of four lights forming the four corners of an imaginary square about a fixation point (see Figure 1). Two response keys were used, one mapped to the top two lights, the other to the bottom two lights. Responding was unimanual. The manipulation crucial to Cotton et al.'s (1977, 1980) argument involved rotating the response keys 90 deg from their standard position parallel to the plane of the subject's body so that they were in line with the subject's midsaggital plane. In this placement, the keys were symmetric with respect to the stimuli occurring to the left and right of fixation, and thus Cotton et al. claimed that no S-R compatibility effects should be anticipated. Nonetheless, a pattern of results was obtained in both Experiment 2 of Cotton et al. (1977) with a l-sec stimulus duration and Experiment 1 of Cotton et al. (1980)with a 100msec stimulus duration that appeared to the authors to be very much like typical S-R compatibility effects. Reaction times were faster to the top light in the visual field ipsilateral to the hand of responding and to the bottom light in the visual field contralateral to the hand of responding. For example, when using the right hand, responses to Light 1 were faster than to Light 2, but responses to Light 3 were faster than to Light 4. These results, as well as a replication