Types of Research Designs

A number of different research designs are used to evaluate treatments and answer other questions about treatment procedures. Each type of design has its own scientific and practical strengths and limitations, and each is ideally suited to answer particular types of questions.

Single-case experimental designs. 

Many studies demonstrating the outcomes obtained with ABA-based procedures use single-case experimental designs (also termed “single-subject designs”; Kazdin, 2010 & 2013) because this type of design is ideal for examining how the behavior of an individual changes as a function of changes in the environment – which is the subject of interest in the field of ABA. These studies often include a small number of individuals (typically, between one to four). It should be noted that published studies using single-case experimental designs are not the same as “case reports” (often seen in clinical journals), which are typically simply descriptive in nature. Rather, studies using single-case designs are controlled studies where treatment is applied in a manner that allows one to demonstrate that the treatment was responsible for the change in behavior. These studies are methodologically rigorous because they involve direct observation of behavior, objective data collection where behaviors are defined and counted (often using a computerized data collection system), and a second observer collects data independently to ensure reliable and accurate data collection.

The most common type of single-case design is a reversal design, which involves the following: a pre-treatment baseline level of behavior is obtained, then treatment is applied and after a change is observed, the treatment is withdrawn, then reapplied to replicate the treatment effect (Kazdin, 2010; Kratochwill & Levin, 2010). The “replication” of the treatment effect illustrates that the treament (and not some other event) is responsible for the change. This type of design has excellent “internal validity” which refers to the extent to which the change in behavior can be attributed to the intervention and not some other variable. Single-case designs are limited, however, in that one cannot determine the extent to which the findings for one study are applicable to other individuals or situations (that is, it has weak “external validity”). It is possible that only cases for which treatment was successful were included in the published study (a concern termed “publication bias”. On the other hand, the ABA literature spans four decades and describes the efficacy of these treatments across a wide range of populations, settings, and problems. Collectively, this extensive body of literature provides strong evidence supporting the external validity of ABA-based interventions.

In the field of ABA, single-case experimental designs are not reserved for exclusive use in research studies. Rather, their use represents good clinical practice. During assessment, single-case designs permit one to identify what factors cause the behavior in question. These findings are then prescriptive for developing an individualized treatment. In addition, single-case designs enable one to determine whether a prescribed treatment (or what particular elements of a treatment) is responsible for behavior change. Isolating the active ingredients of treatment is crucial in saving time and resources. 

Consecutive controlled case series designs. 

Consecutive controlled case-series studies describe a series of cases where single-case experimental designs were used (see Rooker et al., 2013 for a recent example) These studies decribe all individuals encountered who were treated with a certain procedure (regardless of whether the treatment was effective or not), and thus have better external validity than cases involving fewer participants). Because all the cases in the series evaluated treatment using single-case experimental designs, consecutive controlled case-series studies have excellent internal validity as well. Moreover, because a large number of individuals are included, they provide an opportunity to answer other questions, including determining what characteristics predict good outcomes. Several large scale conconsective controlled case series studies describing ABA-based assessment and treatment procedures have been published, and their findings nicely correspond to the broader body of single-case studies decribing smaller numbers of individuals.

Group designs. 

In contrast to single-case experimental designs where the individual’s behavior change during treatment is compared to his/her own behavior without treatment, group designs evaluate treatments based on a comparison of a group of individuals receiving one treatment relative to another similar group of individuals who received no treatment (or a different treatment; Kazdin, 2003). In contrast to single-case designs, where the behaviors of an individual are observed extensively and repeatedly (often for many hours or days) before and after treatment, group designs involve fewer observations of each individual in the group but obtain these measures across large numbers of individuals. Statistical analyses are used to determine whether overall differences between the groups are large enough to conclude that they are not due to normal variation or “chance” (Cohen, Cohen, West, & Aiken, 2003).

The most rigorous type of group design is a randomized controlled trial, which involves randomly assigning participants to a particular group (e.g., treatment or no treatment), and observers who evaluate the outcomes of the treatment do not know whether the participant received treatment or not (i.e., observes are “blinded”). When certain types of treatments, such as medications are being evaluated, the participant may also be “blind” to which group s/he is assigned through the administration of an inactive pill (a placebo). Several group studies describing comprehensive ABA-based interventions for individuals with autism have been published, including some that have used randomization (e.g., Sallows & Graupner, 2005; Smith, Groen,& Wynn, 2000). The most appropriate design to use in a particular situation depends on numerous factors, including the research question, consideration of the relative costs and benefits to participants, and the current state of knowledge about the topic of interest).

Bradley L. Schlaggar, M.D., Ph.D., Named President and CEO of Kennedy Krieger Institute

We’re thrilled to welcome Bradley L. Schlaggar, M.D., Ph.D., to the Kennedy Krieger family as our next President and CEO.

Learn more.

Appointments & Referrals



Read inspiring stories, news and updates about the Institute's patient care, research, special education, professional training, and community programs.


Resource Finder


A free resource that provides access to information and support for individuals and families living with developmental disabilities.