The Autism Society Event and Education Recordings Archive

Some individuals with autism exhibit problem behavior that greatly interferes with their quality of life. Results from numerous studies have shown that these behaviors are learned and can be reduced using treatments based on functional analysis outcomes. Conducting treatment before identifying the function of a behavior problem often leads to ineffective treatments and the need for aversive intervention. This presentation will discuss the importance of conducting a functional analysis, describe how to conduct one, and review appropriate function-based treatments. Individuals with autism may display a variety of challenging behaviors including self-injury, aggression, and destruction. These behaviors often interfere with skill acquisition and impede appropriate social interaction. Research over the past 20 years has shown that these behaviors are learned and are maintained by the consequences that they produce. Challenging behavior may be maintained by social attention. For example, a parent may provide attention in the form of statements of concern (e.g., “Stop hitting yourself”) or physical attention (e.g., hugs the child or blocks self-injury) when his/her child exhibits self-injury. Other challenging behavior may be maintained by escape or avoidance of unpleasant tasks or situations (i.e., the behavior increases because it results in termination of an ongoing activity). For example, a teacher may instruct a student to go to time-out after the student throws his work materials during a math assignment thus allowing the child to escape from the assignment. Challenging behavior may also be maintained by automatic positive reinforcement (i.e., the behavior increases because of the stimulation produced by the response). For example, repetitive finger twirling may be maintained by tactile stimulation to the fingers.

A Functional Analysis involves systematically manipulating environmental events to identify the variable(s) maintaining a participant's problem behavior. Before the development of this methodology, treatment selection for problem behaviors was often arbitrary and required superimposing reinforcers or punishers on top of the reinforcers maintaining problem behavior. Practitioners sometimes prescribe treatments based on the form of a response rather than the potential variables maintaining the response. However, because different problem behaviors (head banging versus hand mouthing) may be maintained by different reinforcers and the same problem behavior may be maintained by different reinforcers for different individuals, selecting treatments based on the form of a response can often lead to ineffective intervention. For example, one individual's self-injury may be maintained by access to attention, whereas another individual's self-injury may be maintained by escape from work-related activities. Reinforcement-based interventions that do not take into account the behavior's maintaining reinforcer are usually ineffective. For this reason, clinicians, parents, and teachers often must resort to aversive consequences to obtain effective treatment outcomes.

Clinicians, teachers, and parents prefer to use nonaversive (i.e., reinforcement-based) interventions. However, use of such interventions requires knowledge about the variables that maintain problem behavior. That is, reinforcement-based interventions require the use of differential reinforcement (delivering the maintaining reinforcer contingent upon an appropriate response) and extinction (withholding the maintaining reinforcer following the problem behavior). In order to implement such procedures, the specific variable(s) maintaining the problem behavior must first be identified. For this reason, a functional analysis approach to assessment and treatment is critical for eliminating the need for aversive approaches to treatment.

A functional analysis provides information about what in the environment maintains an individual's problem behavior. In order to identify the function of a participant's behavior, a number of methods have been developed that fall under the general term “functional assessment.” It is important to note that the terms “functional analysis” and “functional assessment” are not the same. A “functional assessment” can refer to a number of different methods aimed at identifying the variables maintaining problem behavior; however, they differ greatly in terms of their reliability and validity. The simplest approach to functional assessment involves conducting an indirect assessment, which requires the clinician or therapist to ask a person who works closely with the participant a series of questions about the behavior. Although this procedure is efficient and easy to implement, it has been found to be unreliable (different outcomes are obtained across interviewees) and invalid (it does not lead to correct identification of the maintaining reinforcer). For these reasons, an indirect assessment is not recommended as the sole method of conducting a functional assessment. However, it may serve as a useful first step in gathering some general information about the individual and their behavior prior to using a more rigorous form of assessment.

A more systematic approach to assessment involves conducting a descriptive analysis. In a descriptive analysis, the individual's behavior is directly measured under naturalistic conditions. Because the behavior in question is directly measured, it is more objective. However, a descriptive assessment can often be very complex and time consuming. In addition, it only identifies how often certain variables are correlated with the problem behavior rather than identifying the variables maintaining the problem behavior. For example, a descriptive analysis can tell you that attention often follows an individual's problem behavior in the natural environment. However, it does not tell you if this individual's behavior occurs because attention follows the problem behavior. In addition, it is possible that escape may serve as a maintaining reinforcer for an individual's behavior even if it is delivered only rarely following the response in the natural environment. For these reasons, a descriptive analysis is not the recommended for identifying the function of behavior, and interpretations regarding the function of behavior based on this method should be made with caution.

The most rigorous form of assessment is the functional analysis method because it is the only method that allows for the identification of the variables maintaining problem behavior. The model of functional analysis developed by Iwata, Dorsey, Slifer, Bauman, and Richman (1982/1994) has served as the hallmark in the field of applied behavior analysis. A functional analysis involves conducting a series of test conditions in which the potential maintaining variable is delivered following the problem behavior and a control condition in which the potential maintaining variable is continuously provided independent of the problem behavior. Test conditions are designed to detect sensitivity of problem behavior to social positive reinforcement (in the form of attention), social negative reinforcement (in the form of escape from tasks), and automatic reinforcement (no stimulation/toys or attention). Results from Iwata et al. showed that a functional analysis resulted in the detection of a maintaining variable for 6 of the 9 participants in their study. The advantage of this method is that it allows for the identification of functional relations (rather than correlational data). A reported limitation of this method is that it is complex and may have insensitivity to idiosyncratic events. However, recent studies have shown that students and teachers can be quickly trained to implement functional analyses (Iwata et al., 2000; Wallace, Doney, Mintz-Resudek, & Tarbox, 2004). In addition, a number of studies have shown that this method can easily be modified to incorporate idiosyncratic variables (Camp et al., 2000; Fisher, Adelinis, Thompson, Worsdell, & Zarcone, 1998).

This presentation will begin by reviewing what a functional analysis is and why it is important for effective treatment identification. We will then outline three different types of functional assessment procedures, noting advantages and disadvantages of each method. Next, we will review how a functional analysis based on that described by Iwata et al. (1982/1994) is conducted. First, we will review a number of important steps to make prior to initiating a functional analysis (e.g., informed consent, medication, session length, selection of target response, & materials to include in assessment conditions). Next, we will review the purpose and antecedents/consequences provided during each of the standard functional analysis conditions (alone, attention, play, and demand). Following this, we will review how to interpret outcomes obtained during the functional analysis based on visual inspection of the data. A brief review of some potential modifications of functional analysis conditions that can be made to enhance clear outcomes will then be discussed. In addition, practical enhancements (e.g., decreasing session length, decreasing assessment duration, and using different measures for the target response) for increasing the efficiency of the functional analysis method will be reviewed. Next, we will review appropriate reinforcement-based interventions that match the function of the problem behavior identified based on the results of the functional analysis. Finally, the presentation will conclude with a discussion of future directions for enhancing functional analysis procedures for individuals with autism, including how to increase the efficiency and generality of this methodology.

Content Area: Behavior Issues and Supports