Discrimination and generalization in autistic children

The present study examined stimulus control in autistic children. A matching-to-sample procedure was employed in all experiments. In the first part of Experiment I, autistic and control subjects were trained to discriminate between a vertical line and a line tilted at an angle of 33 degrees from vertical. Following training, subjects were given a generalization test to determine the degree of dimensional control by line tilt. In the second part of Experiment I, subjects were trained to discriminate between a vertical line and lines tilted progressively closer to vertical. Experiment II was also a test for the degree of dimensional control by the line tilt. In Experiment I, the autistic subjects took a greater number of trials than the controls to reach the criterion of 24 consecutive correct trials. However, the difference in the number of trials taken by the two groups was not large. There was also little difference between the autistic and control subjects in part two of Experiment I. All of the autistic subjects successfully discriminated between a vertical line and a 2 degree line tilt to a criterion of eight consecutive correct trials. In the generalization tests in Experiments I and II, there was little difference between the autistic and control subjects in dimensional stimulus control. In Experiment III, the autistic subjects were examined for acquisition of a multidimensional discrimination. Both autistic and control subjects were trained to match a standard stimulus with one of four comparison stimuli that were varied in shape and in the presence and absence of a star within the shape. The autistic subjects took a greater number of trials than the controls to reach the criterion of eight consecutive correct trials. However, the difference between the autistic and control subjects in the number of trials taken to reach criterion was not large. In summary, the study found little difference between autistic and control subjects in the acquisition of simple or multidimensional discrimination. As well, there was little difference between the autistics and the controls in dimensional stimulus control. The results of the study suggest that the autistic child's problem is not one of stimulus selectivity.Arts, Faculty ofPsychology, Department ofGraduat