Functional Behavior Assessment for People with Autism. Beth Glasberg

behavior persisted because it helped him avoid his work. If we punish him by sending him to the principal’s office (away from his work), we are inadvertently rewarding his behavior and increasing the chances that he will hit again.

      A final problem with this type of punishment is that an individual may “habituate,” or get used to, a certain level of punishment over time (e.g., Ratner, 1970). This requires the person doing the intervention to gradually increase the intensity of the punishment in order to maintain its effectiveness. For example, the first time that you want your son to stop poking his sister, you may be able to deter him with a stern voice. Over time, you may need to raise your voice a bit. Soon, you may find yourself actually yelling, and so on. Eventually, you may arrive at some level of punishment that you are not comfortable with. When you consider escalation of other forms of punishment, from putting something that tastes bad into a child’s mouth up through spanking, you can see how the need for increased intensity becomes more and more problematic.

       Time-Out

      A second intervention for difficult behaviors that has historically been very popular and remains very popular is “time-out” (Ferster, 1958). There are various permutations of time-out procedures, but the most basic element of this intervention is the removal of an individual from the setting where the behavior problem occurred. For example, a student who hits in class may be removed from the classroom for a given period of time. Time-out offers an alternative to more traditional, physical forms of punishment-based discipline, such as spanking, and is firmly entrenched in current parenting culture. Stop any parent at your local playground and ask how long a time-out for a five-year-old should be. They will likely inform you, without a moment’s hesitation, to use one minute for every year of the child’s age. Parents and teachers everywhere have become time-out experts.

      Despite time-out’s celebrity, there are problems with relying on it as the solution to every behavior problem. In fact, at times, using time-out actually might encourage a child to engage in a behavior more often. For example, in 1994, Brian Iwata and his colleagues summarized their work with over 150 people with developmental disabilities who were hurting themselves, a classic example of a seemingly senseless behavior. The research revealed that almost 40 percent of the time, self-injury was occurring as a way of getting out of or away from something. Like the boy described above who hit to escape work, these individuals were hurting themselves in an effort to leave a situation. As you might imagine, if a child who engaged in self-injury was sent to time-out, then we would see a cycle in which the child is motivated to escape a situation, hurts herself, and escapes the situation.

      More recently, in a sample of thirty-two children seen in an outpatient clinic for varied problem behaviors, a group of researchers (Love et. al, 2009) found that escaping a demand was the second most common motivation for misbehaving. If these children were sent to time out (escape), then we would expect more, rather than less, problem behavior in the future.

      Another challenge associated with time-out is that many children require a great deal of intervention to get them to their time-out location. This intervention may take the form of repeated instructions to go to time-out, hand holding while walking to the time-out area, or more involved physical guidance. Each of these strategies is associated with a high level of attention. If a child is engaging in a problem behavior in order to get attention, time-out may actually lead to rewarding the inappropriate behavior.

       Current Thinking about the Purpose of Challenging Behavior

      In the 1980s, there was a shift in the field of behavioral psychology toward viewing problematic behaviors as a form of communication. To change behaviors, we now focus on the question, “What message is that behavior communicating?” rather than on what the behavior looks like. We now know that even the most seemingly senseless behaviors make sense for the person performing the behavior.

      Problematic behaviors are learned in the same manner as other forms of communication: an individual who has some need identifies a behavior to get his need met. The example above illustrates this: the boy who learned that hitting his instructor resulted in the removal of his work communicated, “Take away my work, please” by hitting. Other consequences, such as getting his hands held down by his sides or getting a talking-to from the principal, do not prevent him from learning this communication. What matters for him at that moment is that he is motivated to escape his work, and, as a result of hitting or swiping materials off of the table, he does.

      Research done by Ted Carr and Mark Durand (1985) was instrumental in developing this clearer understanding of behavior problems. They documented predictable relationships between problem behaviors and the individual’s environment. For example, certain people would have problem behaviors only when they were receiving very little attention. Others might display their problem behaviors whenever they were given a very difficult task to do. Different individuals engaged in their problem behaviors under different conditions. The same behavior in two different individuals might be evoked by two very different sets of conditions. By demonstrating these relationships, Carr and Durand helped to teach us that what a behavior looks like tells us very little about its origin or how to treat it.

      A real-life illustration of Carr and Durand’s work is provided by the behaviors of two teenage boys, Andrew and Robert. Consultation was requested for each of the boys for the same reason, aggression. Andrew blocked the door whenever a one-to-one pull-out teaching session, such as speech or counseling, was ending. He then poked his teacher if she tried to leave the room. If the staff member persisted in attempting to leave, Andrew upped the ante by threatening to stab her with scissors or to hit her with a telephone. Further assessment of this behavior revealed that Andrew behaved this way in order to maintain the rich, one-to-one attention that he had access to during these supportive therapies.

      Like Andrew, Robert also poked people. Whenever his siblings sat near him or walked close enough, a swat or a poke was guaranteed. However, a closer look at Robert’s behavior revealed a story very different from Andrew’s. Robert poked people in order to make them go away. The longer they stayed, the more forceful the poking and swatting became. As you might imagine, Robert’s siblings learned very quickly to stay away from him.

      A particularly exciting result of Carr and Durand’s study was the finding that teaching and reinforcing an appropriate request for whatever was desired weakened the problem behaviors. For example, teaching Andrew to ask the teacher to stay longer resulted in less door-blocking, poking, or threatening with scissors. Teaching Robert to ask for time alone got rid of his poking. The success of this approach, referred to as “functional communication training,” clearly demonstrates that the difficult behaviors so often displayed by individuals with autism spectrum disorders are not an automatic byproduct of the disorder, but rather are learned behaviors, open to treatment using all of the principles of learning.

      Other researchers have experimented not only with conditions that might set the stage for a problem behavior, but also with conditions that they suspected might follow a problem behavior (Iwata, Dorsey, Slifer, Bauman & Richman, 1982). This research helped demonstrate that what happens before a behavior can create a motivation for something and that engaging in a problem behavior can result in obtaining what is wanted. To illustrate, consider the boy in our example above who wanted to escape his work: the teacher approaching his desk with work materials created a motivation for him to be someplace else. Hitting his teacher resulted in him being sent to the principal’s office, thereby avoiding his work. The consequences of the behavior were critical to the development of the behavior problem.

      Brian Iwata and his colleagues (1982) demonstrated this idea empirically. They placed people in scenarios that mimicked these supposedly naturally occurring conditions and measured the effect on their behavior. For example, they might withhold attention from a study participant while she played with toys, but then provide attention when she performed the problem behavior (self-injury). Upon the occurrence of self-injury, the experimenter would immediately return his attention to the girl in the form of social disapproval (e.g., “Don’t do that; you’ll hurt yourself!”). When the consequences for the behavior were presented consistently in this way, the researchers were able to reliably