Les applications de la sensibilité sociale en contexte éducatif

Increasing the appropriate response rate of a target student without producing major drawbacks in the classrooms is a major concern for teachers. Several behavioral interventions have been developed to answer this issue. They often require, however, professional supports and may have the disadvantage to focus only on specific and/or isolated problem behaviors. One way to improve teacher skills in managing student behaviors is through the assessment of their social sensitivity (Forget, 1987 [6]). Social sensitivity quantifies the behavioral adjustment of a subject to contingencies of social reinforcement by means of the generalized matching law (Forget and Rivard, 2010 [3]). It hypothesizes that social attention has an operant function on student behavior and that their adjustment follow the matching law predictions. The purpose of the current study is to illustrate social sensitivity as an analytical tool for identifying different possible treatments in a school context. Two primary school students and their teachers were directly observed in their usual classroom. The students had no diagnosis nor exhibited any problem behavior. They were chosen because, according to their teachers, they showed poor attention to the task. A behavior analyst recorded the naturally occurring rates of appropriate and inappropriate behaviors of the children and the rate of attention given by their teacher. Behaviors and reinforcers were recorded according to an interval sampling method (five seconds of observing followed by five seconds of recording). Social sensitivity accounted for 58% and 82% of the variance. Both subjects demonstrated undermatching and a small preference (bias) toward inappropriate behaviors. Subject 1's sensitivity was .73 (moderate undermatching) whereas subject 2's was .29 (strong undermatching). Biases for both subject were respectively -log(.026) and -log(.08), favouring inappropriate behavior and confirming teachers' perception that children had low attention to the task. Results are similar to other experimental and descriptive studies in humans. High-explained variance suggests that social sensitivity can be applied in the classroom to direct interventions. Based on the obtained matching relations of each subject, two treatments are discussed. Given the results of subject 1, the teacher should be careful in choosing the behavior to reinforce, because the subject has a strong tendency to behave inappropriately. The teacher does not have to increase the absolute number of reinforcers, but rather should distribute them according to a ratio favoring appropriate behaviors. Assuring that reinforcers were contingent to appropriate behavior was also emphasized. According to subject 2's results, the teacher should extinguish, as much as possible, reinforcers to inappropriate behavior and increase social attention to appropriate ones. Given a lower sensitivity to social attention, the teacher should expect a modest modification of behavior. Other types of treatment, such as token economy or based on tangible reinforcers, could be considered in this case. Some limits should be considered. First, the number of subject and the sample size were low. Since social sensitivity relies on a within-subject and molar conceptualisation of behavior, they do not appear as essential factors. Second, the treatments proposed herein were not tested. Effects of the intervention could not have been empirically investigated. It would have brought better support to the model and its efficiency in classrooms. The current study suggests that social sensitivity can be used in school context in order to influence and fine tune behavioral interventions. It seems easier to introduce reluctant teachers to behavioral and cognitive therapies. They require less professional support and would provide autonomy to teachers towards classroom management. Finally, more studies are necessary to assess the therapeutic contribution of social sensitivity; nevertheless, the current results are promising and indicate the need for further examination of the model.