What are two reasons for conducting an abab design rather than an aba design?

4.4.1 Reversal or ABAB design

The reversal design demonstrates the effect of the intervention by alternating the presentation and removal of the program over time. The purpose of the design is to demonstrate a functional relationship between the target behavior and the intervention. The first task in a reversal design is to measure the baseline rate of behavior, which describes the behavior in a normal situation before the intervention. The baseline period (referred to as phase A) is continued until the rate of the response becomes stable.

Then, in the experimental phase (referred to as phase B), the intervention is carried out. This phase continues until the behavior reaches a stable level or diverges clearly from the baseline level. Now the change in behavior is evident, but the cause of change is unclear.

In the reversal phase (phase A) the intervention is withdrawn. The target behavior usually returns to or near the original baseline level. The purpose of the reversal phase is to determine whether the behavior would have remained unchanged if the intervention had not been introduced. When the behavior reverts to the baseline, it is possible to reinstate the intervention (phase B). The design is called the ABAB design because the phases A and B are alternated (Kazdin 1975).

Reversal design

In a reversal design, you run consecutive sessions of the same condition until stable responding is achieved. You then switch to a different condition. If the behavior changes in the predicted direction during the test phase (e.g., aggression is on an increasing trend) and not the control phase (e.g., aggression remains low and stable), the challenging behavior can be attributed to that variable, because it is the only difference between the two conditions. There are many ways you can arrange conditions in a reversal design, but we are going to focus on the ABAB design.

The first thing you do in an ABAB design is try to obtain stable responding during either the control or test condition; whichever you pick would be your “A.” If you were to decide to start with a control condition then you would want to see the challenging behavior occurring at low to zero rates or minimally be on a decreasing trend. Once you achieve a predictable and stable pattern then you would introduce a test condition.

You may also choose to incorporate additional test conditions within the reversal design. If you do, simply add a letter, such as “ABAC,” “A” could be for control, “B” could be a test condition for attention, and “C” could be a test condition for escape.

The strengths of the reversal design are that it is the most convincing demonstration of experimental control. It may accelerate learning through repeated contact with the contingencies. In addition the probability of treatment interference is reduced by maintaining the same contingency across consecutive sessions. A limitation of the reversal design is spending more time in assessment because sessions are conducted consecutively until stable responding is evident. This concern is heightened when assessing dangerous and harmful behaviors.

Baseline logic—reversal designs

Baseline logic must be considered when analyzing data within a reversal design. We use the steady-state strategy to establish a stable pattern of responding so that we can draw more accurate conclusions about data that we collect during subsequent phases. The steady-state strategy entails repeatedly exposing the child or adult to a given condition while trying to eliminate or control any extraneous influences on the behavior. This allows the clinician to obtain a stable pattern of responding prior to introducing the next condition.

The three main components of baseline logic include prediction, verification, and replication. A key component of the scientific process is the development of a hypothesis that is based on reason and observation. The hypothesis the clinician develops marks the beginning of the experiment. The clinician starts the analysis by predicting what he or she thinks will happen with the data. The longer the period in which stable responding is obtained, the better predictive powers of those measures. For example, if aggression in the Play condition has been at a rate of two instances per session then the next session was four instances per session, then two per session, then four per session, and that went on for month, the clinician would likely feel more confident about his or her prediction of what the data will be the next session. If the clinician is basing his or her prediction on only two sessions of data, his or her confidence about what the next data point will look like is going to much lower. This example highlights the relevance of number of data points as it relates to graph interpretation.

In the Play condition, if behavior is maintained by a social function, the clinician would expect that rates of the target challenging behavior would remain low since he or she is providing frequent access to attention and tangibles, while also not introducing any demands. Therefore the clinician would predict that if he or she kept implementing the Play condition, data for the target challenging behavior would remain low and stable.

Verification occurs when the data path remains the same in the subsequent control condition phases, after the independent variable has introduced. In other words the clinician is verifying that the data would have remained at the same level, trend, and variability in the original control phase had the clinician not introduced a new condition (e.g., Attention condition) and kept running control sessions instead. During an FA, verification can occur once the clinician has established stable responding in the Play condition (or control condition) and has introduced a test condition and noticed a change in the level of the data. At this point the clinician cannot claim that the introduction of the independent variable in the test condition led to a change in the level of the data, because some other cooccurring variable may have been present at the time of the change (e.g., perhaps the child or adult became ill when the clinician switched to the test condition, which led to a more frequent display of the target behavior, such as aggression). In order to verify that the data would have remained the same had you remained in the Play (control) condition, the clinician needs to return to the Play (control) condition and observe a similar pattern of data as was observed during the previous Play (control) phase.

Replication provides further evidence that the independent variable is the relevant change in the environment that is responsible for a change in behavior. For example, if data were high in the attention condition during the initial phase of the test condition, and then high again in a second phase of the test condition, the effects of that independent variable will have been replicated.

5.05.5.1 Efficacy

A trademark of ABA is the systematic collection of data on the behaviors that are targeted for change. These data are then graphed to display the impact of the intervention in changing the baseline rates of the behaviors. Time series research designs such as reversal designs and multiple baseline designs are typically used to examine the functional relationship between the intervention and the target behavior (Hersen & Barlow, 1984; Kazdin, 1982). While designs such as these have respectable internal validity and are necessary to demonstrate a functional relationship between the treatment and the observed changes in target behaviors, they are not always used in standard clinical practice. Most practitioners rely upon the simple AB (baseline–treatment) design, hopefully with some form of follow-up. While this design is weak in terms of internal validity in that other plausible hypotheses might be offered to account for the behavior change, it is much easier to implement in standard clinical settings. The design will tell the clinician if the behavior has changed in the desired direction, it is just that the clinician cannot be absolutely certain that the behavior change was caused by the intervention. While demonstrations of functional control can only be demonstrated in individual interventions by the better-quality designs, some authors (Harris & Jensen, 1985) have argued that several replications of AB designs will produce the same strength of evidence of functional control as the more elaborate designs.

Regardless of the type of design, the traditional method for determining the existence of a treatment effect was that of visually analyzing the graphs constructed from the data collection. However, some authors (e.g., Crosbie, 1993) have warned against possible errors with this method, and have argued for the use of time series statistical analyses.

One point that supporters of visual analysis of data have made is that a statistically significant change might occur in the occurrence of the target behavior, but that the change might not be large enough to make a difference to the client's everyday functioning, that is, it might not be clinically significant. This issue of clinical significance of outcomes is part of a general construct that has become known as the social validity of the intervention (Kazdin, 1977; Wolf, 1978). Social validity is considered to have three components. The first of these is the social significance of the goals of the intervention, that is, have the right behaviors been selected as targets of change. The second is the social appropriateness of the intervention strategies, that is, if they are socially acceptable. The third is the social importance of the outcomes, that is, if the outcomes are clinically significant.

Notwithstanding the debate about the relative merits of visual and statistical analysis of data, ABA has been used either exclusively or in combination with other procedures to treat a broad range of clinical problems of children. These have included both externalizing disorders such as conduct disorders (Kazdin, 1987) and attention deficit hyperactivity disorder (Barkley, 1990), and internalizing disorders such as anxiety disorders (Siegel & Ridley-Johnson, 1985) and depression (Kazdin, 1990).

One matter that has received substantial attention in the use of ABA is the extent to which clinical effects are generalized. In order to examine generalization more fully, Drabman, Hammer, and Rosenbaum (1979) developed what they referred to as the generalization map. First, they discriminated among four basic types of generalization, namely generalization across settings, behaviors, subjects, and time. Setting generalization was considered to have occurred if there was change in the target behavior in a setting other than the one in which it was treated. Behavior generalization was considered to have occurred if a behavior not targeted changed along with the target behavior. Subject generalization was said to have occurred if the target behavior changed in an individual other than the child on whom treatment was focused. Finally, time generalization was said to have occurred if a change effected in the target behavior was still in place at a later time. In regard to generalization across time, a distinction is sometimes made between those occasions on which the treatment has been terminated and those on which it is still in place. Time generalization is used when the intervention has ceased, but the term maintenance is used if the treatment is still in place.

Each of the four basic types of generalization could occur alone or in combination with another. For example, generalization across settings and behavior would be said to occur when a nontargeted behavior changed in a setting other than the treatment setting. The four basic types can be combined in sets of two, three, or four to produce a map of 16 classes of generalization. Allen, Tarnowski, Simonian, Elliot, and Drabman (1991) reviewed 904 studies published in 28 journals over a 10 year period and found that only about half reported generalization data.

It may be the case that a lack of generalization is not a problem. Take the case of a child who exhibits problem behaviors in the home, but only at home. If an intervention strategy is put in place at home, and this successfully eliminates the behaviors, the issue of setting generalization is not relevant. If some form of generalization is required, however, it needs to be planned for. The intervention program should incorporate elements that promote the likelihood of generalization occurring. The need to actively program for generalization was initially called for by Stokes and Baer (1977), and more recently Stokes and Osnes (1989) presented a description of several programming tactics grouped into three general categories.

They described the first general category as the exploitation of current functional contingencies. This included tactics such as using naturally occurring reinforcers (and punishers) in the program. New behaviors may have to be taught, but, once learned, they will be maintained by the naturally occurring reinforcers. This will help with the fostering of time generalization. Another tactic in this category is to use naturally occurring reinforcers to systematically reinforce any examples of generalization that occur. If a child is on a program to improve behavior at home, reinforcement of better behavior at school by a teacher will promote setting generalization.

The second of Stokes and Osnes's categories was called training loosely. This refers to designing intervention programs that may begin by being tightly constrained, but move over time to being more diverse. An important tactic in such training is the use of what is referred to as multiple exemplars. A child learning a concept such as dog will be better ably to correctly identify all dogs if taught using a large range of examples of a “dog.” A child undergoing compliance training will display better generalized compliance if taught using a range of instruction types and is reinforced using a range of reinforcement types.

The third of Stokes and Osnes's categories was referred to as the incorporation of functional mediators. A good example of a tactic in this category is seen in social skills training programs for children and adolescents. To increase the probability that skills learned in a clinical setting will be implemented in the community, the child carries a cue card which prompts the critical elements of the response required for a particular social situation.

As was indicated earlier, ABA interventions with children typically involve the triadic model in that behavior change in the child is effected through behavior change in a significant adult in the child's environment. When thinking from within the framework of the triadic model, the issue of generalization is complex because consideration must be given not only to generalization of the child's behavior but also to that of the adult. Sanders and James (1983) identified types of generalization that were relevant to parent behavior when training parents to implement intervention programs with their children. They referred to these as generalization across settings, behaviors, children, and time. Generalization across settings refers to the parent's ability to use the child management skills learned in one setting (e.g., the clinic) in a second setting (e.g., the home). Generalization across behaviors refers to the ability of the parent to use skills learned for one child behavior with a second behavior of the child. Generalization across children refers to the ability of the parent to use skills learned for managing one child for the management of a second child. Generalization across time refers to the parent's ability to continue to use learned skills after the training program has been concluded. These types of generalization of parent behavior are clearly an extension of the types of generalization cited by Drabman et al. (1979) in relation to child behavior.

Within the triadic model of treating children's clinical problems, it must be remembered that generalization of the behavior of the significant other in the child's environment must be planned for. The suggestions offered by Stokes and Osnes (1989) for promotion of generalization apply equally to the behavior of the adults as they do to the children who are the target of intervention. Sanders and Dadds (1993) provide an excellent set of suggestions for promoting the generalization of skills learned by parents in child management training programs.

5.05.7.1 Efficacy

A trademark of ABA is the systematic collection of data on the behaviors that are targeted for change. These data are then graphed to display the impact of the intervention in changing the baseline rates of the behaviors. Time series research designs such as reversal designs and multiple baseline designs are typically used to examine the functional relationship between the intervention and the target behavior (Hersen and Barlow, 1984; Kazdin, 1982). While designs such as these have respectable internal validity and are necessary to demonstrate a functional relationship between the treatment and the observed changes in target behaviors, they are not always used in standard clinical practice. Most practitioners rely upon the simple AB (baseline-treatment) design, hopefully with some form of follow-up. While this design is weak in terms of internal validity in that other plausible hypotheses might be offered to account for the behavior change, it is much easier to implement in standard clinical settings. The design will tell the clinician if the behavior has changed in the desired direction, it is just that the clinician cannot be absolutely certain that the behavior change was caused by the intervention. While demonstrations of functional control can only be demonstrated in individual interventions by the better-quality designs, some authors (Harris and Jensen, 1985) have argued that several replications of AB designs will produce the same strength of evidence of functional control as the more elaborate designs.

Regardless of the type of design, the traditional method for determining a treatment effect is that of visually analyzing the graphs constructed from the data collection. However, some authors (e.g., Crosbie, 1993) have warned against possible errors with this method, and have argued for the use of time series statistical analyses. Recent years have seen increasing use of statistical methods to summarize research in meta-analyses of single case experimental studies (Carr et al., 2015).

One point that supporters of visual analysis of data have made is that sometimes a statistically significant change in the occurrence of the target behavior, might not be large enough to make a difference to the client's everyday functioning, that is, it might not be clinically significant. This issue of clinical significance of outcomes is part of a general construct that has become known as the social validity of the intervention (Kazdin, 1977; Wolf, 1978). Social validity is considered to have three components. The first of these is the social significance of the goals of the intervention, that is, have the right behaviors been selected as targets of change? The second is the social appropriateness of the intervention strategies; are they socially acceptable? The third is the social importance of the outcomes; are the outcomes clinically significant?

Notwithstanding the debate about the relative merits of visual and statistical analysis of data, ABA has been used either exclusively or in combination with other procedures to treat a broad range of clinical problems of children. These have included both externalizing disorders such as conduct disorders (Kazdin, 1987) and attention deficit hyperactivity disorder (Barkley, 1990), and internalizing dis- orders such as anxiety disorders (Siegel and Ridley-Johnson, 1985), and depression (Kazdin, 1990). In recent years the greatest increase in the application of ABA has been in the treatment of Autism Spectrum Disorders (ASD). An overwhelming body of research evidence attests to its effectiveness (e.g. Eldevik et al., 2009; Reichow, 2012; Smith and Iadarola, 2015; Virués-Ortega, 2010) such that it is now the most widely supported evidence-based treatment approach (e.g. National Autism Center, 2015; Wong et al., 2014) and endorsed by the Surgeon General of the United States. In 1999 a report on mental health prepared by the Surgeon General of the United States stated, “Thirty years of research demonstrated the efficacy of applied behavioral methods in reducing inappropriate behavior and in increasing communication, learning, and appropriate social behavior.” (U.S. Department of Health and Human Services, 1999, p. 164).

One matter that has received substantial attention in the use of ABA is the extent to which clinical effects are generalized. In order to examine generalization more fully, Drabman et al. (1979) developed what they referred to as the generalization map. First, they discriminated among four basic types of generalization, namely generalization across settings, behaviors, subjects, and time. Setting generalization was considered to have occurred if there was change in the target behavior in a setting other than the one in which it was treated. Behavior generalization was considered to have occurred if a behavior not targeted changed along with the target behavior. Subject generalization was said to have occurred if the target behavior changed in an individual other than the person on whom treatment was focused. Finally, time generalization was said to have occurred if a change effected in the target behavior was still in place at a later time. Response maintenance is the term now used when the behavior change persists even though all or part of the intervention responsible for the initial behavior change has ceased.

Each of the four basic types of generalization could occur alone or in combination with another. For example, generalization across settings and behavior would be said to occur when a nontargeted behavior changed in a setting other than the treatment setting. The four basic types can be combined in sets of two, three, or four to produce a map of 16 classes of generalization. Sometimes a lack of generalization may not be a problem. Take the case of a child who exhibits problem behaviors in the home, but only at home. If an intervention strategy is put in place at home, and this successfully eliminates the behaviors, the issue of setting generalization is not relevant. If some form of generalization is required however, it needs to be planned for. The intervention program should incorporate elements that promote the likelihood of generalization occurring. The need to actively program for generalization was initially identified by Stokes and Baer (1977). Stokes and Osnes (1989) presented a description of several programming tactics grouped into three general categories.

They described the first general category as the exploitation of current functional contingencies. This included tactics such as using naturally occurring reinforcers (and punishers) in the program. New behaviors may have to be taught, but, once learned, they will be maintained by the naturally occurring reinforcers. This will help with the fostering of response maintenance. Another tactic in this category is to use naturally occurring reinforcers to systematically reinforce any examples of generalization that occur. If a child is on a program to improve behavior at home, reinforcement of better behavior at school by a teacher will promote setting generalization.

The second of Stokes and Osnes's categories they called training loosely. This refers to designing intervention programs that may begin by being tightly constrained, but move over time to being more diverse. An important tactic in such training is the use of multiple exemplars. A child learning a concept such as dog will be better able to correctly identify all dogs if taught using a large range of examples of a “dog.” A child undergoing compliance training will display better generalized compliance if taught using a range of instruction types and if reinforced using a range of reinforcement types.

The third of Stokes and Osnes's categories was referred to as the incorporation of functional mediators. A good example of this tactic is seen in social skills training programs for children and adolescents. To increase the probability that skills learned in a clinical setting will be implemented in the community, the child carries a cue card which prompts the critical elements of the response required for a particular social situation.

ABA interventions with children typically involve a triadic intervention model in that behavior change in the child is effected through behavior change in a significant adult in the child's environment. When thinking from within the framework of this triadic model, the issue of generalization is complex because consideration must be given not only to generalization of the child's behavior but also that of the adult. Sanders and James (1983) identified types of generalization that were relevant to parent behavior when training parents to implement intervention programs with their children. They referred to these as generalization across settings, behaviors, children, and time. Generalization across settings refers to the parent's ability to use the child management skills learned in one setting (e.g., the clinic) in a second setting (e.g., the home). Generalization across behaviors refers to the ability of the parent to use skills learned for one child behavior with a second behavior of the child. Generalization across children refers to the ability of the parent to use skills learned for managing one child for the management of a second child. Generalization across time refers to the parent's ability to continue to use learned skills after the training program has been concluded. Within the triadic model of treating children's clinical problems, it must be remembered that generalization of the behavior of the significant other in the child's environment must be planned for. The suggestions offered by Stokes and Osnes (1989) for promotion of generalization apply equally to the behavior of the adults as they do to the children who are the target of intervention. Sanders and Dadds (1993) provide an excellent set of suggestions for promoting the generalization of skills learned by parents in child management training programs.

Chaser

Vaz et al. (2012) used a chaser to decrease packing of solid foods in three children with feeding disorders. Specifically, the feeder delivered a small bolus of preferred food or liquid following acceptance of a target bite. For two participants, the authors examined the effects of a liquid chaser on packing using an ABAB reversal design (phase A was baseline and phase B was presentation of the chaser). For the third participant, they evaluated packing in the presence and absence of a solid chaser using a multielement design. Across all participants, packing decreased to low levels when the feeder presented a chaser to the participant following acceptance of the target food.

The results of Vaz et al. (2012) are particularly meaningful because the treatment involves a strategy used by typically eating children (e.g., children often use chasers to swallow difficult to swallow food). Despite these benefits, the chaser may only be effective for children who readily accept and swallow certain liquids or solid food (Vaz et al., 2012).

2.1 Within-series Designs

The basic logic and structure of within-series designs are simple; namely, to evaluate changes within a series of data points across time on a single measure or set of related measures (see Hayes et al. 1999). Stability is judged for each data point relative to other data points that immediately precede and follow it. By convention, such designs include comparisons between a naturally occurring state of behavior (denoted by the letter A) and the effects of an imposed manipulation of an independent variable or intervention (denoted by different letters such as B, C, and so on). A simple A-B design involves sampling naturally occurring behavior (A phase), followed by repeated assessment of responding when the independent variable is introduced (B phase). For example, suppose a researcher wanted to determine the effects of rational-emotive therapy on initiating social interactions for a particular client. If an A-B design were chosen, the rate of initiations prior to treatment would be compared with that following treatment in a manner analogous to popular pre-to-post comparisons in-group outcome research. It should be noted, however, that A-B designs are inherently weak in controlling for threats to internal validity (e.g., testing, maturation, history, regression to the mean). Withdrawal/reversal designs control for such threats, and hence provide a more convincing case of experimental control.

Withdrawal designs represent a replication of the basic A-B sequence a second time, with the term withdrawal representing the imposed removal of the active treatment or independent variable (i.e., a return to second baseline or A phase). For example, a simple A-B-A sequence allows one to evaluate treatment effects relative to baseline responding. If an effect due to treatment is present, then it should diminish once treatment is withdrawn and the subject or client is returned to an A phase. Other variations on withdrawal designs include B-A-B designs and A-B-A-B (reversal) designs (see Barlow and Hersen 1984, Hayes et al. 1999). A-B designs and withdrawal/reversal designs are typically used to compare the effects of a finite set of treatment variables with baseline response levels. Data regarding stable response trends are collected across several discrete periods (e.g., time, sessions), wherein the independent variable is either absent (baseline) or present (treatment phase). Phase shifts are data-driven, and response stability determines the next element added to the design, elements that may include other manipulations or treatments either alone (e.g., A-B-A-C-A-B-A) or in combination (e.g., A-B-A-B+C-A-B-A). As the manipulated behavior change is repeatedly demonstrated and replicated across increasing numbers of phase shifts and time, confidence in the role of the independent variable as a cause of such changes increases.

This basic logic of within-series single-subject methodology has been expanded in sophisticated and at times complex ways to meet basic and applied purposes. For instance, such designs can be used to test the differential effects of more than one treatment. Other reversal designs (i.e., B-C-B-C) involve the comparison of differential, yet consecutive, treatment interventions across multiple phase shifts. These designs are similar to the above reversals in how control is evaluated, but differ primarily in that baseline phases are not required. Designs are also available that combine features of withdrawal and reversal. For example, A-B-A-B-C-B designs allow one to compare A and B phases with each other (reversal; A-B-A-B), and a second treatment to B (reversal; e.g., B-C-B), or even to evaluate the extent to which behavior tracks a specified behavioral criterion (i.e., changing criterion designs; see Hayes et al. 1999). Changing criterion designs provide an alternative method for experimentally analyzing behavioral effects without subsequent treatment withdrawal. Criterion is set such that optimal levels given exposure can be met and then systematically increased (or decreased) to instill greater demand on acquiring new repertoires. For example, a child learning to add may be required to calculate 4 of 10 problems correctly to earn a prize. Once the child successfully and consistently demonstrates this level of responding, the demand increases to 6 of 10 correct problems. Changing criterion designs serve as a medium to demonstrate learning through achieving successive approximations of the end-state.

Single-Case Evaluation Designs

Single-case evaluation methodology is a mainstay of ABA research (Kazdin, 2011) and the basis of many sports related studies (Luiselli, 2011; Martin et al., 2004). The publications we reviewed in this chapter are testimony to the variety of single-case designs available to researchers. Of course, these designs are intended to control for internal validity through time series analysis and replication of effects through experimental manipulation of independent variables. Repeated evaluation of procedures that produce similar outcomes contributes to the external validity of single-case studies.

Briefly, reversal designs have value in quickly evaluating stimulus control and antecedent interventions (Osborne et al., 1990) but may not fit well with consequence-based interventions targeting skill acquisition (i.e., learned skills may not “reverse”). Multiple baseline designs are particularly adaptable, in part because of versatility. For example, the multiple baseline design across behaviors adequately meets the common coaching objective of training more than one skill in a single athlete (Boyer et al., 2009; Brobst & Ward, 2002; Ward & Carnes, 2002). For team sports, interventions can be evaluated efficiently in a multiple baseline design across players (Harrison & Pyles, 2013; Kladopoulos & McComas, 2001). In a changing criterion design, the steps within a task analysis naturally serve as the criteria for measuring the effects of shaping procedures (Scott et al., 1997). One additional strategy, the alternating treatments design, makes it possible to compare two or more procedures (Wolko et al., 1993), thereby enhancing coaching and training efficiency.

Independent of research, single-case evaluation designs should have considerable appeal to practitioners. Whether training athletes directly or consulting with coaches, professionals working in the sports arena can employ single-case designs to evaluate learning trends and make necessary procedural revisions that improve performance. Tkachuk et al. (2003) also commented that single-case methodology can be used to demonstrate “to the consumers of sport psychology services that measurable improvements in athletic performance are due to the interventions” (p. 112). Their suggestion is consistent with Gee's (2010) assertion that applied sport psychology research plays a vital role in educating athletes and coaches about the mechanisms by which professional consultation can positively influence performance.

Behavioral Assessment

The methods used to collect data on occurrence of human stereotypy have drawn recent attention from researchers. Although studies have evaluated stereotypy using a variety of indirect (e.g., rating scales) and direct (e.g., rate or duration) measures, the percentage of time engaged in the stereotyped response is typically the dependent measure of interest. The percentage of time measure is determined by collecting data on the duration of time an individual engages in stereotypy within an observation period and dividing that number by the total number of seconds of the observation period and multiplying by 100%. This method is sometimes referred to as continuous duration recording (CDR). Given the labor intensity of this method of data collection, many clinicians and researchers prefer to use time sampling or interval methods to evaluate stereotypy.

One time-sampling method that generates very accurate estimates of the duration of events such as stereotypy is momentary time sampling (MTS). This method involves breaking an observation period into equal intervals and recording the occurrence or nonoccurrence of the behavior during the last second of each interval. For example, a 10 min (600 s) observation period can be divided into 60 intervals of 10 s. Using 10 s MTS to score stereotypy, an observer scores the occurrence of target behavior during only the last second of each 10-s interval. The number of intervals with the occurrence of the target behavior is divided by the total number of intervals and then multiplied by 100% to arrive at the percentage of intervals the individual engaged in the target behavior. As a whole, the literature suggests that data collected using 10-s MTS are comparable to data collected with CDR.

In behavior analytic studies, data on specific forms of stereotypy are individually collected and the effects of environmental events (e.g., potential interventions) are typically evaluated using single-subject experimental designs such as an ABAB reversal design or a multielement design. An ABAB reversal involves two conditions wherein ‘A’ is a baseline or no-intervention (or no manipulation) condition and ‘B’ is an intervention condition. Observations of a specified length (e.g., 15 min) are first conducted under the baseline condition. After a minimum of three observation sessions are conducted, and relatively stable levels of behavior are observed, the intervention phase is implemented. This phase is also conducted for a minimum of three sessions and until behavior levels stabilize. Once stable behavior is observed, the intervention phase is withdrawn and the process is repeated. A multielement design is similar to a reversal design except that a baseline condition is typically compared to one or more interventions denoted ‘B’, ‘C’, ‘D’, and so on. In addition, instead of conducting the same condition until stability is achieved, the conditions are rapidly alternated in a random or quasi-random order.

Operant interpretations of stereotypy are derived, at least in part, from data obtained via functional analysis methodologies, which were pioneered by Iwata and colleagues in the early 1980s through the mid-1990s. Functional analysis procedures that are used to evaluate the operant function of stereotypy involve systematic manipulations of both antecedents (e.g., presence of a task) and consequences (e.g., attention from a parent). Manipulations are made within specific environmental conditions to isolate events that may be correlated with the presence or absence of stereotypy. In that light, functional analysis represents a broad spectrum approach to identifying behavioral function. Much like an allergist applies many different sample materials to an individual’s skin to assess his sensitivity to potential allergens, functional analysis places behavior in a variety of contexts to assess the individual’s sensitivity to potential reinforcers. The effects of different environmental conditions are evaluated using single-subject experimental designs similar to those described above. Numerous studies involving a variety of problem behaviors (e.g., self-injurious behavior, aggressive behavior, habit behavior) have shown that data collected via functional analysis methodologies lead to the development of more effective behavioral interventions than interventions that are developed without such assessment.

Why is an ABAB design better than an AB design?

What are the benefits of using an ABAB treatment design?

What is the purpose of an ABAB design?

What is the difference between ABA and ABAB design?