The locus of stimulus probability effect in choice reaction time

Abstract

A choice reaction time experiment was conducted in which subjects were to respond with one response to either of two stimuli, S1 and S2, and with another response to a third stimulus, S3. The stimuli in the low-compatibility condition were three colors and the responses were letter names. In a high-compatibility condition, S1 and 82 were upper-and lowercases of the same letter and S3 was another letter. The responses were the letter names. The relative frequency of S1 and S2 was manipulated. The results showed a Compatibility by Relative Frequency interaction, with relative frequency affecting reaction time only under low compatibility. It was argued that the stimulus relative frequency affects the response selection stage of information processing.Using a 1 to I stimulus-response (S·R) mapping in a choice reaction time (CRT) task typically shows reaction time to increase with a decrease in the relative frequency of the stimulus. This fact has been shown to hold, at least when S·R compatibility is not particularly high. The increase in reaction time (RT) as a function of decrease in relative frequency is obtained if relative frequency is varied by manipulating the number of equally likely alternatives (Hick , 1952). It is also found when the probability of occurrence of particular stimuli is manipulated within a set of n possible alternatives (Hyman, 1953).There are at least two possible explanations of this phenomenon . The subject might be responding faster with a more probable response , or he might encode faster the more frequent stimulus. If a I to I S·R mapping is used , these two possible interpretations cannot be separated.In an ingenious study, laBerge and Tweedy (1964) attempted to isolate stimulus probability from response probability . They used three stimuli (SI, S2, and S3) and two responses (Ra and Rb), such that Sl and S2 were associated with Ra and S3 was associated with Rb. With this technique , variation in the relative frequency of the SI and S2 should result in RT data that is analyzable with respect to the response vs. stimulus frequency inte rpretations of the relative frequency effect. Specifically , as laBerge and Tweedy argue, if the relative frequency does not affect the stimulus processing stage, then the mean RTs for Sl and S2 should not differ. laBerge and Tweedy used three colors as stimuli and two finger pressing responses and found that the more frequent of SI and S2 was responded to faster. They concluded, therefore , that the speed of stimulus processing is affected by stimulus frequency .There seems, however, to be a third possible explanations of the relative frequency effect. Response bias (the term used by laBerge and Tweedy) could refer to two Dean Purcell sponsors thiS paper and takes full editorial responsibility for its contents. distinct stages in the information processing system . One is the response selection stage, the other is the response execution stage. The laBerge and Tweedy design allows variation in stimulus probability independently of response execution pr...

Case specificity of the stimulus probability effect

“…Within the larger context of the entire information processing system, stimulus probability is emerging as a variable whose influence is seen at many stages. It has been shown that probability has a larger effect when stimulus-response compatibility is low, for example, indicating an effect of probability on some phase of response selection as well as on perceptual mechanisms (Hawkins et al, 1973;Spector & Lyons, 1976). In view of this diversity of established probability effects, it seems inevitable that any complete model for probability effects must address multiple processing stages.…”

Effects of stimulus–response probability on choice reaction time: Evidence from the lateralized readiness potential.

“…Despite considerable research into this possibility, however, there is still disagreement about whether perceptual adjustments do in fact contribute to the probability effect (e.g., Blackman, 1972;F. De Jong & Sanders, 1986;Dykes & Pascal, 1981;Gescheider, Wright, Weber, Kirchner, & Milligan, 1969;Miller & Pachella, 1973;Orenstein, 1970;Sanders, 1977Sanders, , 1980Spector & Lyons, 1976). Effects of perceptual manipulations like stimulus intensity and stimulus discriminability sometimes have no larger effects for low-probability stimuli than for high-probability ones, contrary to what would be expected if probability affected the rate of perceptual processing.…”

Expectancy and frequency effects on perceptual and motor systems in choice reaction time

“…This finding suggests that probability influences the time required by perceptual processes to recognize a stimulus. Similarly, the effect of S-R compatibility is generally larger for low-probability stimuli than for high-probability stimuli (e.g., Hawkins et al, 1973;Sanders, 1970;Spector & Lyons, 1976), suggesting that probability also affects the time needed for response selection. Indeed, Blackman (1975) obtained interactions of probability with both stimulus quality and S-R compatibility in a single experiment, although these variables were additive with each other.…”