Effects of stimulus similarity on the magnitude of behavioral contrast

Abstract

Two experiments were performed in order to collect data which would contribute to a resolution of the competition of various theories of behavioral contrast. Experiment I investigated the amount of contrast produced by two extremes of stimulus similarity. A mixed schedule was used for the discrimination between a maximally similar set of stimuli and a multiple schedule for the discrimination between a maximally different set of stimuli. The major findings of this experiment was that the magnitude of behavioral contrast is inversely related to the degree of stimulus similarity between the positive and negative stimuli, suggesting that contrast is an integral part of the discrimination process. Experiment II allowed a comparison of the aversive properties of the two negative stimuli used in Experiment I as well as an evaluation of the birds' preference for the various components used in Experiment I. Aversiveness was measured by the number of time outs earned by a subject during the various components of Experiment I. Experiment II found that the negative component in which there was the greatest response suppression, not the most errors, was the most aversive component. Also, assuming that relative preference is the inverse of relative aversiveness, Experiment II found that the component in which the response rate increase was greatest was the most preferred and that the preference was greater for both components in which contrast occurred than it was for the negative components. Of the four theories of contrast, only Terrace's inhibitory control hypothesis and the preference hypothesis of Bloomfield can account for the findings of these two experiments. Since Reynolds' hypothesis says nothing of the function of the stimulus in the production of contrast, its adequacy in accounting for contrast must be questioned. Furthermore, the role that Amsel suggests primary frustration plays in the production of contrast must also be questioned since Amsel also predicts no relationship between the magnitude of behavioral contrast and stimulus similarity and since contrast was not seen to increase with the aversiveness of the preceding nonreinforced component.