Behavior is composed of physical events. Applied behavior analysts must make obvious the relevant quantitative characteristics of those physical events. Some of these characteristics include the following (Martin & Pear, 1998). The topography of a behavior refers its specific muscular movements. For example, Martin & Pear have described the proper form of above-water arm movement when swimming the back stroke as: (1) fingers together, (2) roll shoulder into your ear, (3) arm comes over straight, (4) arm comes over close to ear, and (5) little finger enters water first. The amount of a behavior refers to its frequency or duration. For example, we might count the number of aggressive episodes throughout the day by a client at a group home and/or how long each episode lasts. The intensity of a behavior refers to such measures as force or loudness. For example, we might measure how hard a tennis player can serve the ball. The latency of behavior refers to the time between a stimulus presentation and the onset of the behavior. For example, how long does it take before a child complies to a parental request?
Often the applied researcher does not have option of choosing a response that can be easily recorded through mechanization. Instead, he or she is required to use human observers to quantify the behavior of other human beings. A relevant question to consider, then, is not was the target behavior changed, but whose? Baer et al. (1987) state that "applied behavior most often still is, and most often always should be, the study of an observer's behavior that has been brought under the tight control of the subject's behavior" (p. 317). This study entails modifying the observer's behavior through careful direct training and subjecting it to direct and frequent reliability assessments. The key is to increase the correspondence between what the observer recorded and what actually happened.
A written code determines what the observer records. If the code writer is the person who finds the behavior to be problematic, then satisfying the code writer "is the only approach to valid measurement that makes sense" (Baer et al., 1987, p. 316). This suggests a strong argument against standard codes - a code determined by the complaint of one person's behavior is unlikely to satisfy the complaint of someone else's behavior.
Applied behavioral research "usually studies what subjects can be brought to do rather than what they can be brought to say" (Baer et al., 1968, p. 93). A program that results in a client reporting that he is no longer smoking, when in fact he has not stopped, would appear to have little applied value. For this reason, self-reports are considered "risky." This is not to say that direct observation by an outside observer is without bias, but such bias can be assessed and altered to a much greater degree than that of the client's self-observation.
Baer et al. (1987) note that in recent years applied behavior analysis has been mainstreamed with other disciplines, such as "behavior therapy, education, developmental psychology, psycholinguistics and sociobiology." Consequently, applied behavioral researchers have been confronted with constructs such as "anxiety, attention, intelligence, disabilities, spontaneity, readiness, critical periods, innate releasers, storage and retrieval mechanisms, schemata, and the like." They advise that the most fruitful task is "to recognize that each of these labels (and many others like them) often represents some behavioral reality not yet analyzed as such" (p. 315). For an interesting recent example of how the anxiety construct has been behavioralized, see Friman, Hayes, & Wilson (1998).
Nonexample
Reports by classroom teachers, therapists, and teachers indicated that autistic children attending biweekly therapy sessions were unmotivated and engaged in frequent self-stimulatory behavior (e.g., rocking back and forth). It was agreed that a reasonable goal was to increase correct task responding from each child's academic curriculum (e.g., "touch your nose" versus "touch my nose"). Previous research suggested that identifying reinforcers for autistic children can be difficult, in that these children often do not respond to stimuli that interest other children (e.g., toys) or to social reinforcers (e.g., praise). The researchers were aware of the Premack Principle, which states that the opportunity to engage in a behavior that occurs frequently can be used to reinforce a behavior that occurs less often. With this in mind, the intervention included prompting the children to engage in 3-5 seconds of self-stimulation following correct task responding. When teaching the "touch your [body part]" versus "touch my [body part]" discrimination, the teacher was responsible for presenting the instruction, prompting the self-stimulation when a correct response occurred, and recording the data. The teacher counted a correct response when the child pointed in the general direction of her body part or the teacher's body part when asked to do so.
Example
Reports by classroom teachers, therapists, and teachers indicated that autistic children attending biweekly therapy sessions were unmotivated and engaged in frequent self-stimulatory behavior (e.g., rocking back and forth). It was agreed that a reasonable goal was to increase correct task responding from each child's academic curriculum (e.g., "touch your nose" versus "touch my nose"). Previous research suggested that identifying reinforcers for autistic children can be difficult, in that these children often do not respond to stimuli that interest other children (e.g., toys) or to social reinforcers (e.g., praise). The researchers were aware of the Premack Principle, which states that the opportunity to engage in a behavior that occurs frequently can be used to reinforce a behavior that occurs less often. With this in mind, the intervention included prompting the children to engage in 3-5 seconds of self-stimulation following correct task responding. When teaching the "touch your [body part]" versus "touch my [body part]" discrimination, the teacher was responsible for presenting the instruction, prompting the self-stimulation when a correct response occurred, and recording the data. A teacher counted a correct response when the child's finger made contact with either her body part or the teacher's body part when asked to point to it. A second independent observer recorded the child's correct responses based on this same criterion.
Analysis
The first item is not behavioral in at least two respects. First, the physical event ("pointing in the general direction") that constitutes a correct response is imprecisely defined. It may be difficult on some trials to determine exactly what it is that the child is pointing to. Second, there is no procedure in place to determine the reliability of the teacher's recording. If she had a bias either for or against the child, then her recorded data may not be an accurate reflection of the child's actual performance.
The procedure in the second item attempts to avoid these two behavioral problems. First, the target behavior is more precisely defined ("finger makes contact with body part"), leading to less ambiguity about what response is and is not correct. Second, by having a second observer keep data, the researchers could compare these records with the teacher's and assess the amount of agreement. If it is high, then we have greater confidence that the data accurately reflect what transpired during that training session; if it is low, then perhaps one of the observers is biased or perhaps the response definition requires further refinement.
Related Source: Charlop, Kurtz, & Casey (1990)