ABSTRACT
The effect of different relationships between stimuli and responses, and of number of stimuli, on choice—reaction-times was investigated using an apparatus in which the stimuli consisted of vibrations delivered to the tips of the fingers through the keys used for making responses. The tasks involved either pressing the key which vibrated (”compatible”) or one of the keys of the unstimulated hand (”mapped”). In some sessions all relationships between stimulus and responses were compatible. In other sessions, all were mapped, and in yet others half were compatible and half mapped. The fully compatible task was a replication of Leonard’s (1959) experiment, which gave no increase of latency with degree of choice. This finding was supported, and evidence is given which suggests that these compatible latencies are at some minimum level. However, in the mapped conditions substantial increases of latency with degree of choice were found. In the mixed conditions the latencies of responses to the compatible stimuli were dependent on degree of choice, and increased logarithmically relative to the pure compatible conditions. The latencies to mapped stimuli, however, were found to decrease relative to the pure mapped case. The results raise difficulties for previous models of choice—reaction time. A new model is suggested which fits the results obtained here, and those of previous choice—reaction experiments. It also reconciles the conflicting formulations of Hick (1952 a) and Hyman (1953), showing both to be limiting conditions of a general case.
