To begin our discussion about common intervention for individuals with autism it is important to first define and discuss Applied Behavior Analysis (ABA). Many times the term ABA is used, incorrectly, to describe the intervention being implemented (i.e. “I am doing ABA intervention.”). ABA isn’t an intervention, it is the application of the science of learning to increase or decrease socially significant behaviors. There are, however, interventions based on the science of ABA which address a variety of socially significant behaviors. It is also important to note here that ABA is NOT and Autism-specific science. Actually, the field of ABA is based on the experimental analysis of behavior work of B.F. Skinner who developed the concept of Operant Conditioning. Most of his original research was conducted with animals. His research involving human subjects culminated in the book (and techniques) “Verbal Behavior” which will be discussed later this month.
To provide a framework for this audience, I’m going to base my description of ABA on the seminal article published in 1968 by Donald M. Baer, Montrose M. Wolf and Todd R. Risley. In this article, the authors described 7 terms which can be used to evaluate and intervention or a study to determine whether or not it is truly reflective of Applied Behavior Analysis. These seven terms are: applied, behavioral, analytic, technological, conceptually systematic, effective and generality. For sake of this discussion, we will focusing on discussing these terms in relation to interventions. Interventions based on the science of ABA need to follow these guidelines:
The intervention must be applied to a behavior which is socially significant. Socially significant behaviors are those which are important for an individual to do (or not do) in order to successfully function within their environment (i.e. home, school, community, etc.). These behaviors can include, but are not limited to, communication, social skills, eating, sleeping, physical skills, and disruptive behaviors. If the skill being addressed by an intervention has social significance to the individual and their community, then the intervention fits this component.
For an intervention to be considered behavioral the behaviors being addressed must be clearly defined in observable and measurable terms. When a behavior is clearly defined in a way it can be seen and recorded, then the changes in behavior can be measured over time. If we are able to do this, we are able to see progress over time which will allow us to determine whether or not the intervention being implemented is effectively changing behavior.
To show the intervention is responsible for behavior change, the intervention must be analytic. This means that data need to be collected on a regular basis (i.e. daily or weekly) to determine whether or not the intervention, as it is being implemented, is effective in improving the target behavior. If the data is not changing in the predicted way (i.e. increasing or decreasing) then the intervention may need to be changed in some way. This point is extremely important because if data is not collected on a regular basis, and behavior change is not effectively monitored, an intervention may be being implemented which has no impact on behavior and is therefore a waste of time and money.
An intervention must be technological in the sense that the techniques being implemented need to be completely identified and described. If the techniques are not clearly defined and described, then the likelihood that person A will implement the intervention the same as person B is very, very low. This component is extremely important when considering procedural fidelity and it’s impact on behavior change. This means if the data is not showing behavioral improvement, then there are two things to consider: (1) the intervention may not be effective for the individual, or (2) the intervention is not being correctly implemented (i.e. the fidelity of implementation is low). If the intervention is technologically describe and the people implementing the intervention are trained to correctly implement the procedures, then the likelihood of the intervention being implemented correctly across interventionists is much higher.
To be conceptually systematic an intervention needs to utilize strategies which are relevant to specific principles defined within the body of research around behavior analysis. These principles include, but are not limited to, positive and negative reinforcement, positive and negative punishment, extinction, fading, error-less discrimination, differential reinforcement, prompting, and chaining. When the technological description of an intervention describes the intervention procedures in terms of behavioral principle being utilized the intervention are much easier to learn and teach.
To determine whether or not an intervention is effective depends less on whether or not there was behavior change and more on whether or not the extent of behavior change is socially significant. To illustrate this I’ll give you two examples. When implementing an intervention to teach a child to use expressive labels to request items (i.e. say “milk” to get milk) if the intervention only increases the number of labels used as requests to 5 over the course of a year then the intervention produced change but this change may not be seen as significant. This is because, presumably, there are more than 5 things the individual needs or wants to request. Another example has to do with hitting. If an intervention is implemented to decrease the frequency of hitting behavior and the behavior decreases to zero with adults at home but is still occurring at school with adults and peers, then the intervention has not effectively changed behavior. In both of these cases there was behavior change, but the extent of behavior change is the concern. If the change in behavior is happening too slow, there may be something about the intervention that needs to be changed or the intervention may need to be changed altogether.
The final thing to consider when assessing interventions is to look at the generality of behavior change. This means that we need to look at whether or not the behavior changes persist over time and generalize to other environments. This is important because behavior change is only significant if it maintains over time and the individual is able to use the skill in a variety of environments. If the individual does not automatically generalize the skills being learned, then this needs to be an added component of the intervention. The only way to determine whether or not the skills are generalizing is to take data across environments.
In the next few weeks we will be discussing a variety of interventions based on the science of ABA. These interventions are used to increase and decrease a variety of target behaviors, have been studied in the research community and applied to a variety of individuals. Our goal during this month is to educate you on effective interventions as well as teach you how to make informed decisions about beginning, continuing and discontinuing interventions. Please feel free to ask questions about any intervention.