Stimulus generalization and anxiety in schizophrenic and normal subjects

Sixty schizophrenics and 60 normal subjects were compared on a test of stimulus generalization. The generalization stimuli were lines differing in length. Both schizophrenics and normals were divided into two groups (30 subjects each) and trained either on a short line or on a long line. In addition to the generalization stimuli, subjects were presented with an irrelevant stimulus (circle). Although there was a significant difference between the generalization gradients of the groups, this difference was limited to generalization stimuli away from the negative training stimulus. The concept of peak shift was invoked to interpret these data. That the groups did not respond to the irrelevant stimulus is inconsistent with interference theory and gives some support to the differentiation between conceptual overinclusion and perceptual overinclusion. The relationship between stimulus generalization and scores on the Manifest Anxiety Scale was inconsistent.