ABAB DESIGN The ABAB design is one of the oldest and most widely used single- case designs developed in behavioral psychology. It was initially used in laboratory studies with animals (Sidman, 1960); however, as the applied behavior analysis movement got under way (Baer, Wolf, & Risley, 1968), it became a prototype for applied behavioral investigations conducted in the natural environment. Although the number of single- case designs has increased markedly since the early days of applied behavior analysis (e.g., Kazdin, 1980; Kratochwill, 1978), the ABAB design still occupies a prominent place in applied behavioral research. Moreover, because of the high degree of experimental control that it provides, it has been widely used with individuals manifesting various types of handicaps (Bergan, 1977). For example, the ABAB design has been particularly useful in studying environmental variables affecting language acquisition in retarded children (Bergan, 1977). The ABAB design is intended to reveal a functional relationship between an experimental treatment and a behavior targeted for change. For example, it might be used to establish a functional relationship between the use of the plural form of a noun and a treatment such as praise following the occurrence of a plural noun. The demonstration of a functional relationship between praise and plural nouns would require an association between the frequency of plural- noun production and the occurrence of verbal praise. Given that a functional relationship were established, verbal praise could be assumed to function as a positive reinforcer increasing the probability of occurrence of plural nouns by the subject or subjects participating in the experiment. The ABAB technique has often been referred to as a single- case design (e.g., Kratochwill, 1978). However, it may be applied with more than one subject. Thus, the term single case is a bit misleading. Glass, Wilson, and Gottman (1975) among others called attention to the fact that the ABAB design is a time- series design in that it refl ects an effort to determine changes in behavior occurring across a series of points in time. Recognition of the ABAB design as a time- series design opened the way for linking the design to the statistical procedures associated with time- series analysis (see, for example, Glass, Wilson, & Gottmann, 1975). Application of time- series analysis procedures affords a statistical test for hypotheses that may be investigated with the ABAB design. However, despite this advantage, time- series techniques have not been widely used in applied investigations involving the ABAB design. There are a variety of reasons for this. Among them is the fact that the graphing techniques suggested by behavioral psychologists (e.g., Parsonson & Baer, 1978) as an alternative to statistical analysis are easier to implement and to interpret than time- series statistics. Nonetheless, time- series procedures constitute a potentially powerful tool for applied behavioral research and their use can be expected to increase in the future. As the letters in its name suggest, the ABAB design includes four phases. The initial A phase is a baseline period that records behavior across a series of points in time in the absence of intervention. The length of the baseline period varies depending on the variability of the behavior being recorded. If the behavior is highly variable, a longer baseline is required than if the behavior is highly stable. More data are required to get a sense of the fl uctuations that may be expected without intervention for a highly variable behavior than for a highly stable behavior. The second phase, denoted by the letter B, is a treatment phase. During this phase the treatment is introduced. The treatment may be implemented in accordance with a variety of different schedules. For example, treatment may be implemented with every occurrence of the target behavior. For instance, praise might be given following every occurrence of a plural noun. On the other hand, treatment might be implemented in accordance with one of the many available partial reinforcement schedules. Thus, praise might be given after every third occurrence of a plural noun. The third phase, also denoted by the letter A, constitutes a return to baseline. The return to baseline may be brought about by various means. One is to withdraw the treatment. For instance, praise might not be given following plural- noun utterances during the return- to- baseline phase. Another procedure is to introduce another treatment intended to bring the target behavior back to baseline level. For example, reinforcement of a behavior that is incompatible with the target behavior may be introduced during the return- to- baseline phase. The fi nal phase in the ABAB design, denoted by the second occurrence of the letter B, is a second implementation of the treatment. The second implementation is intended to demonstrate treatment control over the target behavior by minimizing the possibility that environmental infl uences occurring coincidentally with the treatment could be responsible for the observed behavior change. The major advantage of the ABAB design lies in the fact that it minimizes the likelihood of coincidental environmen-tal infl uences on the target behavior. There are two potential disadvantages to the approach (Kazdin, 1973). One is that some behaviors are not easily reversed. For example, a skill that has been well- learned may not be easy to unlearn. The second disadvantage is that there are cases in which it may not be practical to carry out a return- to- baseline even if it is possible to do so. For instance, a teacher may not want to return a child’s performance of an academic skill to baseline even for a short period of time. Despite these shortcomings, the ABAB design has been shown to be useful in establishing a functional relationship between a treatment and behavior in countless applications. It is truly a mainstay in applied behavioral research and will continue to be used widely.