No abstract

Selecting the correct response from a set of alternative responses is typically a time-consuming cognitive operation. This fundamental aspect of human performance is reflected in the almost universal finding that simple reaction times (RTs) are faster than choice RTs. A generalization of the effects of response uncertainty on RTs is given by Hick's law, which states that the mean RT is proportional to the logarithm of the number of response alternatives (Hick, 1952;Schweickert, 1993). However, differences between choice RTs and simple RTs may not be entirely attributable to the time course of response selection; response selection is not the only component process required in the performance of choice RT tasks not required by simple RT tasks. Choice RTs also require stimulus identification before the correct response can be determined, whereas simple RTs only require detection of the imperative stimulus before a prespecified response can be initiated. Kveraga, Boucher, and Hughes (2002) obtained a set of results for which it was likely that response selection processes were entirely responsible for increases in mean RT that accompanied increases in the number of equally likely response alternatives. In one experimental condition, Kveraga et al. measured saccadic RTs when the number of potential target locations varied from 1 to 8 in octave steps (one, two, four, and eight target locations). This location uncertainty produced a corresponding response uncertainty. The level of stimulus-response (S-R) uncertainty was constant within a block of trials but varied between blocks of trials. Circular annuli indicated the possible target locations within any one block of trials. When one of these annuli was filled with a luminous disk, subjects were instructed to look at that location as quickly and accurately as possible. The results showed that saccadic RTs were completely unaffected by the number of potential target locations (see also Saslow, 1967). Not only did saccadic RTs violate Hick's law, they also violated the virtually universal finding that choice RTs are longer than simple RTs. Kveraga et al. also examined the effects of S-R uncertainty in a keypress task using identical stimulus conditions. In this second task, each target location was associated with a different key. S-R uncertainty varied across blocks of trials just as it had for the saccade task, and the potential target locations were again indicated by circular annuli. The results from the keypress task conformed to Hick's law and included a substantial difference between simple and choice RTs. The effects of S-R uncertainty observed with the keypress task cannot be attributed to processes associated with detection or identification of the targets, since these variables had no effect on saccades using an identical display. Therefore, both tasks required the same target detection and identification components, and both required a different response be associated with each individual stimulus. The tasks differed only in the modality of the response. Kv...