CHAPTER 2

BEHAVIOR THERAPY

Richard E. Zinbarg

James W. Griffith

The central defining feature of behavior therapy is that it involves the application of the laws of learning theory to the modification of problematic behavior. At the time that behavior therapy was first beginning to gain momentum, another aspect that distinguished it from the prevailing Freudian school of therapy was the recognition of the need to distinguish etiological factors from maintaining ones. This distinction no longer can be claimed as being unique to behavior therapy because it is also central to cognitive therapy; however, this distinction still lies at the heart of behavior therapy. We wish to emphasize from the outset that behavior therapy has (not surprisingly) evolved significantly since its inception. Perhaps one of the two most fundamental ways in which behavior therapy has changed is its integration with cognitive therapy such that these days there are probably many more therapists who identify themselves as cognitive-behaviorists (as we do) than behaviorists. The other most fundamental change is that there is more recognition now of the importance of the therapeutic relationship than there was in the beginning.

Whereas contemporary cognitive-behavioral therapy (CBT) almost always combines the behavioral and cognitive traditions, this chapter focuses on the behavioral tradition. Chapter 3 by Scott Kellogg and Jeffrey Young provides an overview of the cognitive tradition. Similarly, Arnold Lazarus, one of the pioneers of the integration of the behavioral and cognitive traditions, writes about multimodal therapy, his approach to CBT, in Chapter 13.

THEORETICAL CONCEPTS OF BEHAVIOR THERAPY

Key theoretical concepts in behavior therapy are classical conditioning, instrumental conditioning, generalization, extinction, and functional assessment. Behavior therapy also gave rise to a key methodological concept: the single case design in psychotherapy.

Classical conditioning is the learning that occurs when two stimuli are presented in close temporal proximity and with some degree of contingency or correlation between them. A form of classical conditioning that is particularly important for behavior therapy occurs when the first of these stimuli (the conditioned stimulus) is affectively neutral prior to being paired with the second (the unconditioned stimulus), whereas the second already has some emotional valence for the individual and elicits some form of response (the unconditioned response) indicative of this valence. In this case, the contingency between the conditioned stimulus and the unconditioned stimulus typically leads to the conditional stimulus acquiring at least some of the emotional valence of the unconditioned stimulus and the capacity to elicit a response (the conditioned emotional response). The conditioned emotional response sometimes resembles the unconditioned response, but does not necessarily need to do so and often does not. As discussed in more detail later, Pavlov (1927) is often credited with being the father of the modern study of classical conditioning, and Watson (1994/1913) and his student Mary Cover Jones (1924) are usually credited for pioneering the application of classical conditioning to the understanding and remediation of problem behaviors in humans.

Instrumental conditioning is the learning that occurs when a response (the instrumental response) is consistently followed by either a positive or a negative consequence (the reinforcer). As stated in Thorndike’s (1927) law of effect, the tendency to emit a behavior that is followed by positive consequences will strengthen a response, whereas the tendency to emit a behavior that is followed by negative consequences will typically weaken a response. Often, the very same behavior that is rewarded in one situation will not be rewarded in a different situation. The cues that help to signal when the different contingencies involving a given response are in effect are called discriminative stimuli because they help the individual discriminate when the different contingencies are in effect. Thus, discriminative stimuli that signal that a given response will be rewarded are called reward cues and elicit hope; those that signal a response will be punished are called passive avoidance cues and elicit anxiety; those that signal that a response will lead to the omission of an otherwise expected reward are called omission cues and elicit frustration; and those that signal that a response will lead to the omission of an otherwise expected punishment are called active avoidance cues and are thought to elicit anxiety (on early learning trials) and relief (on later learning trials).

Generalization refers to the notion that when one conditioned or discriminative stimulus comes to elicit an emotional response, other stimuli that resemble the conditioned or discriminative stimulus will also elicit some of this same emotional response. It is thought that there is a gradient of generalization such that the more similar stimuli are to the conditioned or discriminative stimulus, the stronger the generalized responding will be to them.

Extinction involves the associative changes that occur when the unconditioned stimulus or reinforcer is no longer contingent on the conditioned stimulus, instrumental response, or discriminative stimulus. In the case of a conditioned stimulus, presenting it in the absence of an unconditioned stimulus leads to a decrement in the strength of the conditioned response. An instrumental response that previously had been rewarded will similarly experience a decrement in strength when no longer followed by reward. Similarly, the emotions conditioned to a discriminative stimulus will weaken when the discriminative stimulus no longer signals that a reinforcer is contingent on responding.

Functional assessment involves the attempt to determine the contingencies maintaining problematic behavior and the discriminative stimuli that signal when those contingencies are in effect and thus serve to elicit the problem behaviors. Early behavior therapists favored functional assessment and eschewed diagnosis and diagnostic labels. Contemporary behavior therapists still emphasize the importance of functional assessment but are more open to diagnoses and diagnostic labels.

Single case designs involve the systematic assessment of problem behaviors or treatment targets across both baseline conditions and conditions in which an intervention or therapeutic contingency is actively implemented. The most simple design is one in which a baseline phase (A) is followed by treatment (B), usually referred to as an A/B design. This is the minimal design necessary to determine that meaningful change in the problem behavior has indeed occurred while also being able to rule out at least some threats to internal validity such as testing, instrumentation, and regression to the mean (Kazdin, 1981). If the intervention is one that does not produce permanent change, then the phase change can be repeated like an A/B/A design that begins with a baseline phase (A) followed by treatment (B) that is followed by a second baseline (A) in which treatment is withdrawn (or an A/B/A/B design that adds a second treatment phase to the A/B/A design). If the problem behavior tracks the phase changes, our confidence increases that the treatment is accounting for the changes. Another common design that is more powerful than the simple A/B design for ruling out threats to internal validity is the multiple baseline design. In a multiple baseline design across behaviors, a researcher measures several behaviors targeted for change and applies treatment to them sequentially while continuing baseline measurement on the behaviors not yet targeted. If four behaviors are targeted for change, the design begins with collecting baseline data on all four, and then the treatment is applied to the first behavior while the baseline is continued for the other three behaviors. Subsequently, the treatment is applied to the second behavior while the baseline is continued for the other two behaviors and so on. If each behavior remains relatively stable throughout its baseline and does not begin to show improvement until sometime shortly after the treatment is applied to it, we can rule out additional threats to internal validity and increase our confidence that the treatment caused the observed changes. (For an excellent discussion of multiple baseline designs across behaviors as well as across individuals, see Barlow & Hersen, 1984.)

HISTORY OF BEHAVIOR THERAPY

The history of behavior therapy is closely linked to the history of learning theory in psychology and the rise of behaviorism that took place in the early twentieth century. Thus, an understanding of the history of behavior therapy requires some understanding of the history of learning theory and the role of key historical figures. Indeed, one strength of behavior therapy is its close relationship to basic behavioral science in human and nonhuman animals. The historical connection between the development of learning theories and behavior therapy is evident in a historical time line published by The Behavior Therapist (vol. 29, 2006). Some important people in the history of learning theory are Ivan P. Pavlov, James B. Watson, B. F. Skinner, and Orval H. Mowrer. Their roles in the development of learning theory are discussed in the following sections, as well as how their work relates to contemporary behavior therapy.

Ivan Petrovich Pavlov (1849–1936)

Pavlov was a...