I recall sitting in my psychology statistics class years ago, stressing over experimental types and the complex math involved in calculating p-values. All of the focus in my undergraduate statistics classes was on group experimental designs, and professors typically glazed over single-case designs (SCEDs), because they were “not as common” and that we weren’t likely to use them in the future.
Imagine my surprise when I began my master’s program and we immediately dove into the importance of SCEDs in the field of ABA. SCEDs play an incredibly important role in applied behavior analysis and provide an empirically supported way to evaluate the effects of interventions on an individual’s behavior.
By focusing on only one subject or one small group of subjects, behavior analysts can uncover valuable findings that are applicable in clinical, educational, and therapeutic settings. This article will go a little deeper into this “not as common” experimental design, with some deeper focus on ABA and ABAB designs. They have some similarities, but differ in meaningful ways.
What Is a Single-Case Design in ABA?
A single-case experimental design is a research method that is used in ABA. It involves repeated measurements of a subject’s behavior across different phases. These phases typically involve baseline and intervention phases.
The goal is to observe changes in behavior over time and over phases to determine if the changes in behavior are due to the interventions being introduced or if changes are due to external factors. These designs are very flexible, can be applied in real-world settings, are effective for smaller groups than group designs, and can help determine if there is a functional relationship.
It is important to know that “single-case” does not necessarily mean that it is limited to one person. Rather, it means that the individual serves as their own control. For group designs, there is often a control group that is separate from the experimental group – one does not get the intervention (known as the baseline group), while the other group does get the intervention (the experimental group).
In single-case studies, the subject goes through a baseline phase(s) and experimental phase(s). Comparing a behavior before, during, and after an intervention can help to establish a functional relationship between the intervention (the independent variable of the study) and the behavior (the dependent variable). Showing experimental control (showing that changes in a behavior reliably follow changes in an intervention) is one of the main points of SCEDs and shows that an intervention is meaningful and can make a difference.
Obtén terapia ABA en casa, con cariño
Completa el formulario para conversar sobre cómo tu hijo puede comenzar la terapia rápidamente y sin estrés:
- Therapists Available
- Help With Insurance
- Quick Response
Understanding the ABA Design
The ABA design is a type of SCED that includes three distinct phases: a baseline phase (A), an intervention phase (B), and then a return to baseline phase (A). In an ABA design, a behavior is observed without any type of treatment, with the introduction of an intervention, and then one more time after the intervention is discontinued. The purpose of this type of design is to determine whether the intervention affects the behavior of the subject, more specifically by seeing if the behavior returns to its typical (or baseline) levels after the intervention is removed.
An ABA design can be useful in demonstrating that there is a correlation between an intervention and behavior change. Because it requires fewer phases than other, more extended designs, it is relatively quick to implement and analyze. For example, if a child who is frequently off task starts a token board system and their off-task behavior is reduced, if the behavior increases again once the token system is removed, this shows that there is a potential relationship between the behavior and the intervention. If a behavior change is observable when the intervention is removed or added, it can be assumed that the intervention played a role in that change.
One limitation of the ABA design is that it does not include a replication of the effect, which is to again go into the intervention to see if a change occurs once again (more on that later). Like with group designs, repetition of results means stronger evidence. So, more phases with continuous behavior shifts mean that the evidence for a functional relationship is even stronger.
Another main limitation of this type of design is that sometimes, behaviors do not go back to their baseline levels. This typically happens if the intervention teaches a skill that the learner did not previously have. For example, if a subject hits when they want access to a preferred item, an intervention would likely include teaching them an alternative way to mand for the item. This learning of a new behavior cannot be unlearned, and so reverting to baseline may not see any type of change in behavior from the intervention phase.
Understanding the ABAB Design
The ABAB design, as you might have guessed, adds a B to the ABA design. That is, it goes into a second intervention phase. An ABAB design includes an initial baseline phase (A), an intervention phase (B), withdrawal of intervention (A), and then back to intervention once again (B). The main purpose of this design is to replicate the treatment effect and further strengthen the evidence that a functional relationship between the intervention and behavior exists.
One of the main strengths of the ABAB design is that it can better demonstrate that a functional relationship exists through replication. Remember, more repetition means stronger evidence. This further replication improves that design’s internal validity. For example, the ABA example earlier with the token board suggests that there is a functional relationship, but it is always possible that these shifts in behavior were because of other variables. To better prove that the token board makes a difference, reintroducing the token board should make the behavior change again. The behavior change after a phase change makes it less likely that the behavior change is random.
ABAB designs also have limitations. The biggest concern with ABAB designs, and ABA designs for that matter, is that sometimes withdrawing a successful treatment is unethical. Take, for example, a child with self-injurious behaviors. If an intervention helps to reduce self-injury in a child, withdrawing the intervention will possibly cause them to continue to hurt themselves at a higher rate.
Reducing harm in this case is much more important than further proving a functional relationship exists. Another possible limitation is that it takes more time, effort, and resources than designs with fewer phase changes. Sometimes, especially with more intense behaviors, it’s more manageable to choose a design with fewer phases.
ABA vs ABAB: Key Differences
While both ABA and ABAB designs are used to evaluate behavior changes in applied settings, they do have a few key differences. The most significant difference is the replication step. ABAB designs implement the intervention one more time, but the ABA design does not. As a result, ABA designs do not determine a functional relationship as well as an ABAB design. ABA designs are able to show a correlation, but without the replication step, a functional relationship cannot be confirmed.
Even with ABAB being the stronger of the two for determining a relationship between independent and dependent variables, there are a few reasons why a practitioner may choose an ABA design over an ABAB. For example, ABA designs may be favored during the early stages of intervention development when time or resources are limited. ABAB designs may be preferred if the evaluator has the time and if they need stronger evidence for decision making.
When to Use ABA vs. ABAB in Clinical Practice
Choosing between these two designs depends on several factors, including the severity of behavior, the level of risk to the client, the setting, and goals of the intervention. ABAB designs are a good choice for practitioners who need stronger evidence that an intervention is effective. Sometimes, with analyzing a graph, the behavior change isn’t a large, dramatic shift, and so another replication can help to see if the shift was random or if it was truly because of a useful intervention.
In situations where time and resources are more limited, or when the goal is to gather preliminary data, an ABA design may be more practical. For more severe behaviors, it is recommended that other types of designs be used, such as a simple AB design (not returning to A, keeping the intervention in place).
Common Misconceptions
One common misconception is with the term “ABA design” itself. Personally, my mind goes right into applied behavior analysis when I see ABA. It is easy to assume that ABA refers to the entire field of behavior analysis, but in this instance, it refers to the design type. ABA as a field is much, much broader than this type of specific design.
Additionally, it’s important to know that there are many, many other types of SCED designs. There are more than just ABA and ABAB. There are other letters you can add, more A’s and B’s, other formats like multiple baselines and changing criterion, the combinations are endless.
Finally, as mentioned earlier, not all behaviors will return to baseline. It is important to think about the behavior, the chosen intervention, and the skills required for the chosen intervention before choosing a design that returns to baseline. If it is assumed that the behavior may not return to baseline, it would be a better idea to choose one of the other types of designs.
Tips for Aspiring RBTs and BCBAs
For those entering the field, learning how to collect reliable baseline and intervention data is one of the most important steps to master. Accurate data is arguably the most vital part of good analysis and, therefore, decision making on interventions and treatment.
I also cannot stress enough the importance of considering the ethical implications of any design chosen. Will returning to baseline be safe for the subject? Is keeping a subject in the baseline phase too long something to be concerned about? Always, always ask yourself these types of questions before taking action. In supervision meetings or in treatment planning, be prepared to explain how and why the chosen design supports the individual’s or group’s goals.
It is also a good idea to spend time practicing and getting comfortable with the visual analysis of graphs. Much of working in ABA involves looking at and analyzing graphs, so don’t be afraid to look at graphs quite often.
Finally, for those studying for the RBT or BCBA exam, remember that the A in all SCED designs are baseline phases, and letters B-Z are interventions. Looking at a list of multiple-choice answers that look like a bunch of jumbled-up letters could be intimidating, but remembering that A is baseline and all others are different interventions can help make them a bit less daunting.
Mi punto de vista como BCBA
It is definitely important for BCBAs to understand how these designs work before implementing a behavior plan. If the basics for single-case designs are not understood, implementing plans could very easily set the individual up for failure. A big component of applied behavior analysis is its individualized focus and specifically tailored interventions. What works for some may not work for others. So, choosing a band-aid intervention without testing the intervention, documentation, and visual analysis of graphs is not the best course of action.
It is not very often that I find myself using ABA and ABAB designs, but there have certainly been instances in my career so far. I’d argue that knowing about these (and other) types of designs is just as important for choosing an intervention. BCBAs do not pull these plans out of thin air. Choosing a solid intervention may require research and reading through empirical papers and journals. Being able to read these papers and understand the graphs in them is a required skill.
Conclusión
ABA and ABAB designs differ in structure, strength of evidence, and ethical considerations. ABA designs offer good insights with considerations for time and resource efficiency. ABAB designs go a bit further and take more time, but are much more of a solid choice for establishing a functional relationship between an intervention and behavior. It is ultimately up to the practitioner to choose which one is best for their specific situation.
