Article Summary:
Charlop M., Kurtz P., & Casey F. (1990).
Using Aberrant Behaviors as Reinforcers for Autistic Children. Journal of Applied Behavioral Analysis 23, 163-181.
Goals and Methods
The Premack Principle states that high-frequency activities can be used to reinforce a low-frequency response. When subjects do not respond to stimuli that most find reinforcing (food, social reinforcers), it may be beneficial to identify high-frequency activities, however unusual, and to use these to increase responding. Studies in the past have demonstrated use of aberrant behaviors as reinforcement to increase responding in children with autism. This current study methodically compares responding rates in the presence of a typical reinforcer (food) vs an aberrent reinforcer (opportunity to engage in aberrant behaviors). The researchers hypothesize that the aberrant reinforcer will produce the highest levels of responding. This finding would have important implications for increasing motivation and learning in otherwise unmotivated children, specifically children with autism. This study is divided into three experiments. Each experiment compares the efficacy of food versus the efficacy of a different aberrant Aso, in each study the children were observed for "side-effects" - general increases in off-task and aberrant behaviors as a result of the program. The breakdown is as follows:
Experiment I - Four children participated in experiment I. Levels of responding were compared under four conditions; baseline, food as reinforcer, stereotypy as reinforcer, and varied (stereotypy and/or food.).
Experiment II - Three children participated in experiment II. Baseline levels of responding were compared with three conditions: food as reinforcer, delayed echolalia as reinforcer, and varied (sometimes food, and sometimes echolalia).
Experiment III Three children participated in experiment III. Baseline levels of responding were compared to responding under three conditions: reinforcement with food, with perseverations, and with stereotypy.
Each of the experiments was carried out in a similar setting and with a similar population of children. The children chosen were of grade-school age and suffered from autism. They were described as generally unmotivated and not unresponsive to social reinforcers. The children attended an after-school behavioral program twice a week, where teaching was structured according to traditional behavioral concepts and devices (tasks taught in series of trials, using prompts, reinforcements, corrective trials, etc). The experimental sessions were incorporated into this program without changing either the basic structure of the program or the individual goals for each child. A baseline was taken on each child, during which the usual trials were carried out and only food and praise were given in response to correct answers. After taking baselines of varying lengths for different children, the children were then exposed to different conditions of reinforcement. The authors describe their study design as "multielement". On alternate sessions different reinforcers were offered, either food , the aberrant activity, or a varied offering. There were never more then three sessions in a row offering the same conditions of reinforcement. The data was recorded on a graph for each child, the independent variable being the type of reinforcer used, the dependent variable being the percentage of on-task behaviors. Two observers recorded on task, off-task and aberrant behaviors (in all experiments interrater reliabliity was over 90%).
Post session observations were also made to see if the interventions resulted in behavior change during other activities. The process of choosing reinforcers was done individually for each child. For one child preferred stereotypy was tapping, for another sniffing a fruit-scented marker. Echolalia was also individual; one child liked to say "ding ding ding, you win again," another, "red, red red." For perseverative reinforcement, a preferred object (a humpty-dumpty doll, leaves and cacti, etc.) was made available, and the child was allowed to play with and talk about it.
Results
In all of the experiments, abberant behaviors were shown to be the most potent reinforcers, and undesirable side effects did not surface.
Experiment I - Baseline percentages for on-task behavior in all four children were between 50% and 60%. During intervention sessions, the highest percentage of on-task responding was during the sessions in which stereotypy was used as a reinforcer (levels rose to around 70% and 80%). This was true for all four children. The varied condition was also always at least slighty higher then the food condition. Response levels in the presence of food were lowest. Two of the children's levels acually dipped below baseline.
Experiment II - Baseline levels of responding for all three children were between 55-60%. For two children echolalia was the most potent reinforcer, with average response levels rising to between 70% and 90%. For one of the children the varied reinforcer was slightly higher then the echolalia alone. Food was the weakest reinforcer for all of the children; in one response levels dropped beneath baseline.Experiment III - Baseline levels of responding averaged around 50% and 60%. For all three children, perseverative behavior was the most effective reinforcer, with average levels rising to over 80%. For two of the children stereoptypy was only slightly less effective. For all three, food was the least effective. In all of the experiments, post session observations were done in various settings to observe for any changes in off-task behaviors and levels of aberrant behavior. In none of the experiments were there undesirable side-effects. In experiment I, one of the children was observed for agression, and following treatment, his agression levels actually dropped.
Discussion: implications and generalizations
This article has implications both for further study and practice in the area of reinforcement choice, and also in the handling and understanding of aberrant behaviors in children with autism.
Narrowing the concept of "reinforcement" has negative effects upon learning. While using the wrong reinforcer can create the image of an unmotivated child who doesn't learn, identifying and using preferred reinforcements will increase responding. Educators should carefully examine the effectiveness of the reinforcers they choose, and should not automatically rule out a reinfrorcer because it is bizarre.
This has special implications for the population of children with autism. Children with autism are notorious for not prefering social reinforcers. While food may be reinforcing, when it is the only reinforcer used satiation may result. This was demonstrated by the structure of the current study where baseline data was collected using food as a reinforcer, and then throughout the experimental sessions food continued to be used as one of the conditions. During these experimental sessions, some of the subjects performance levels dropped to below baseline, presumably because satiation had by then occured.
Despite these benefits there are limits to the practice of using abberant behaviors as reinforcement. Not all reinforcers are easily provided by the experimenter. Some, such as eye blinking, are not in the experimenter's control. Others such as watching a traffic light change color can not be brought into the classroom. Still others are so distasteful to others, or even dangerous and destructive (such as self-injurious behaviors) that they cannot be used in good conscience.
This article shows that aberrant behaviors which are merely odd may have uses as effective reinforcers. Equally important, the study systematically examines the "side-effects" of this practise. Past studies show that abberant behaviors are difficult to extinguish, yet the focus of educators has still mainly been to discourage them as much as possible. Fears that the aberrant behaviors would increase if ever encouraged, kept teachers from using them as reinforcements. This study symtematically examined the issue of side-effects, and found the fear to be unwarranted - at least in these cases. Even during experimental sessions, the children were amiable in parting with objects of perseveration (one boy for instance, exclaimed "I love that cow," as he agreeably handed it back to the experimenter). This study allayed many concerns, suggesting that educators should focus on directing and extracting the reinforcing benefits from aberrant behaviors rather then attempting to extinguish them.