The Autism Society Event and Education Recordings Archive

Some individuals with autism require reinforcement to learn new skills. Although the technology exists to identify effective reinforcers, few people are familiar with these techniques. Additionally, many family members, educators, and clinicians use reinforcement systems without consideration of the variables that may impact of the effectiveness of these systems. This presentation will focus on what we know about preference and reinforcement when working with individuals with autism, what we think we know, and what we still need to learn. Many people have negative connotations about using reinforcement with individuals with autism. Some people consider reinforcement to be a bribe. Others feel that it should not be necessary to use reinforcement when working with individuals with autism; they feel that these individuals should be intrinsically motivated. In a perfect world, natural or intrinsic reinforcers would be sufficient to motivate individuals with autism to work and learn, to fulfill their potential. However, for some individuals with autism, in order to maximize their potential, additional reinforcement must be used.

Reinforcement is not a four-letter word. It is one of the most basic and crucial mechanisms in learning, and because it is so basic, we often take it for granted. It's like driving a car—it is so engrained in our repertoire, that we sometimes do it without even thinking about it. We sometimes take reinforcement for granted as well, because it is so basic to everything we do. The reality is, reinforcement governs all our behavior—but, because it is so basic, we rarely think about it. Most of us don't consider our paychecks a “bribe”, and despite what our employers might think, many people are not “intrinsically” motivated to go to work. We work because we receive reinforcement for doing so. For all of us, reinforcement really is crucial for maintaining behavior and learning new skills.

It is well known that some individuals with autism require reinforcement to learn new skills, maintain work behavior, and to decrease challenging behavior. Over the past 20 years, there has been a great deal of published literature on developing techniques to identify efficient and effective reinforcers for individuals with autism. Although applied researchers frequently utilize systematic preference assessments prior to implementing treatment, few parents, teachers, clinicians, and practitioners in applied settings do this.

Research has demonstrated that conducting systematic preference assessments is the most effective way to identify potential reinforcers for individuals with autism. How do most parents, teachers, and clinicians determine the preferences of the individuals they work with? How do they identify potential reinforcers that will be used to maintain work behavior, decrease challenging behavior, and teach new skills? We recently completed a survey that asked this very question. We found that most practitioners assess potential reinforcers simply by asking the individual with special needs what they like; if that individual is nonverbal, the parents and caregivers are asked what the individual prefers. Is this a problem? For some individuals with autism, the answer could be, “Yes”.

One published study addressed the question of whether simply asking caregivers what a child preferred would identify reinforcers as reliably as a systematic preference assessment, which had already been proven to be reliable. Adults were asked to rank which items they thought would be most preferred by the individuals; this was compared to a systematic assessment, where the potential reinforcers were actually given to the children, and their preferences were directly measured. It turned out that there was little correspondence between what the adults thought were reinforcers and what actually functioned as reinforcers. This demonstrated that sometimes simply asking adults was not enough! Thus, the answer to the question, is asking parents, teachers, or caregivers an effective way to identify reinforcers for individuals with autism, the answer is, “Maybe, maybe not”.

The other technique that is most frequently used to assess preferences in individuals with autism is simply to ask the individuals themselves what they prefer. Surely, this must be an effective way to identify reinforcers! Again, the answer is, “Maybe, maybe not”. One recent study asked individuals to rate how much they liked various items. Then, each of these items was tested to see if they functioned as reinforcers. It turned out that asking the individual to identify their own reinforcers was successful in identifying true reinforcers in less than 60% of cases. We have developed slightly different protocols to assess the effectiveness of having individuals verbally identify their preferences, and we have found that in about 30% of cases, simply asking an individual with autism which item he or she prefers did not identify effective reinforcers.

Although behavior analysts have done a wonderful job conducting research on conducting preference and reinforcement assessments, the fact that most parents, teachers, practitioners, and caregivers do not use these techniques suggests that this research has not been translated into practice. Why is it so crucial to identify effective reinforcers for individuals with autism? Because for many such individuals, there will be little learning without effective reinforcers!

This presentation will begin by reviewing what reinforcement is, how it works, and the difference between preference and reinforcement. Recent literature on the technology of assessing preferences and reinforcers will be reviewed. We will also review a new model for conducting preference assessments. The model is basically a large flowchart--at each step in the flowchart, you answer a one-sentence question that assesses the skill of the individual, and based upon the answer, the flowchart points you toward a preference assessment that is likely efficient and effective for that individual. Next, the presentation reviews practical issues related to assessing preferences, such as how often preferences may change, and how motivating operations may influence preference assessment outcomes. Finally, the presentation concludes with some thoughts about what we still need to know about effectively implementing reinforcement systems with individuals with autism, including how to increase the effectiveness of reinforcers in various settings.

Content Area: Behavior Issues and Supports