Variable criterion analysis of brightness effects in simple reaction time.

Abstract

Using light onset as the stimulus in simple reaction time (SRT), the effect of stimulus intensity was studied in both between-subjects and within-subjects experimental designs. There was a strong intensity effect in both conditions but no significant interaction between the effect of stimulus intensity and the type of design. This differs from previous results with auditory stimuli where such an interaction has been demonstrated. When the criterion parameters of variable criterion theory were evaluated directly, the only significant effect was greater criterion variability in the between-subjects condition. Theoretical functions describing the growth of sensory strength for each intensity had different starting points and were largely parallel, showing only late temporal divergence. This provides an explanation of the rarity, in the SRT literature, of interactions between visual intensity and criterion variables. Correlations illustrating the relations between reaction time (RT) measures and theoretical criterion parameters are presented. Absence of the predicted relation between intensity and RT variability is evidence against theories relating RT to impulse rate treated as a Poisson process.

Behavioral estimates of interhemispheric transmission time and the signal detection method: A reappraisal

“…First, the introduction of catch trials is described very well by a change of the detection criterion, as can be concluded from the interaction between stimulus intensity and percentage of catch trials in Experiments I and 2. There has been evidence suggesting that the interaction is more difficult to obtain with visual stimuli than with auditory stimuli (Grice, Nullmeyer, & Schnizlein, 1979;Nissen, 1977), but this does not seem to hold for the manipulation realized here (see also Hawkins et al, 1988). The introduction of catch trials has the effect that more information must be gathered before a response is initiated, as could already be concluded from the percentage of anticipatory reactions and false alarms in the different conditions (see the Results sections).…”

“…Since the original formulation of the variable-criterion theory (Grice, 1968), more elaborate versions ofthe model have been devised (e.g., Grice, Nullmeyer, & Schnizlein, 1979;Grice, Nullmeyer, & Spiker, 1982) in which information does not accumulate linearly from a fixed intercept. However, because these alternative models do not add information to the problem under investigation and because they require elaborate mathematical expressions that burden the treatise, they will not be dealt with in the present manuscript.…”

Stimulus intensity effects between and within subjects in auditory reaction time: A variable criterion analysis

“…Since Grice's (1968) initial presentation, methods have been further developed; a more complete analysis is now made possible by obtaining estimates of the criterion parameters for each subject. Such a reanalysis of the Grice and Hunter data is now especially desirable, since Grice, Nullmeyer, and Schnizlein (1979) failed to find a similar interaction in simple RT for visual intensity. The Grice and Hunter data are not ideal for such an analysis, since the number of trials was rather small (50 to each stimulus in the within condition and 100 in the between condition).…”

“…However, reasonable parameter estimates can be obtained, and the results of the analysis are presented here.' Details of methods for estimating the mean (C) and standard deviations (a) of the criterion distributions for individual subjects have been presented elsewhere (Grice, Nullmeyer , & Spiker, 1977;Grice et al, 1979) . The first step was to obtain estimates of the functions for the growth of sensory strength or detection information (V) from the more reliable (100-trial) data of the soft-only (40 dB) and loud-only (100 dB) groups.…”

“…The Grice and Hunter data are not ideal for such an analysis, since the number of trials was rather small (50 to each stimulus in the within condition and 100 in the between condition). The numbers used by Grice et al (1979) were 120 and 240. However, reasonable parameter estimates can be obtained, and the results of the analysis are presented here.'…”