TRAINING SOCIAL INFERENCE WITH MENTALLY RETARDED ADOLESCENTS

JUDITH L GOODSTONE, Purdue University

Abstract

Mentally retarded individuals are characterized as having maladaptive behavior. Utilizing the model prescribed by Greenspan, social inference may be defined as the comprehension of social cues whose interpretation indicates the probabilities of action in a situation. Educators currently view cognitive strategy training as a viable technique to promote needed social competencies and enhance generalizability in a variety of academic, vocational, and interpersonal settings. Thus, the generalization of social inference by mentally retarded individuals may be enhanced through cognitive strategy training aimed at improving the discrimination of salient cues and the interpretation of their meaning. The research questions asked if training in social inference improved scores on the Test of Social Inference and if the number of training sessions was related to such improvement. The cognitive training technique used was higher-level questions. The design included a training condition where subjects were trained to improve their social inference through prompt questions, and a no-training condition where subjects were exposed to the training items but were not given prompts. Within each condition, there were three groups which differed only in the number of sessions received: one, three, or six sessions. An additional no-exposure to training materials group received only the dependent measure. The main ANOVA approached significant (p (LESSTHEQ) .06) for the main effect of condition. The main effect of number of sessions and the interaction of condition X number of sessions was not significant. Investigation of the variance of TSI scores between cells revealed high homogeneity of variance. There were no significant differences on range tests between groups of means with respect to TSI scores. A one-tailed t-test between means on the TSI for training condition and no-training condition subjects was significant. Descriptive data revealed 73% of the training condition subjects received TSI scores above the grand mean whereas only 37% of the no-training condition subjects received scores above the grand mean. This study may be interpreted as implying that higher-level questions improve social inference. Suggestions for future research are included.

Degree

Ph.D.

Subject Area

Special education

COinS