Multiple Baseline Single-Case Experimental Design

Single-Case Experimental Designs

SCEDs are unique in that they conduct experimental investigations with a single case or series of cases. A case might be an individual or a larger single unit like a business or school. These designs contrast common experimental designs (e.g., randomized controlled trials) that are conducted using large sample sizes, multiple groups (typically including a control group), and few conditions (i.e., one control condition and one experimental condition). Such large nomothetic studies are often considered the gold standard of scientific research. However, they may show differences between study groups because of a small number of skewed results; thus, statistical significance does not always translate into clinically or practically meaningful results. Furthermore, aggregating data within study groups obscures individual variability, which can be relevant to understanding the effects of an intervention on a specific person. By conducting idiographic research with small sample sizes while still allowing for experimental manipulation (e.g., randomization, control conditions) and rigorous data collection and analysis, SCED is able to provide another methodology through which to pursue research questions.

SCEDs also allow for intensive clinical observation (e.g., four children in a behavioral learning condition under frequent observation by the researcher as opposed to 100 subjects providing data at specific time points) that provides researchers with insights into individual differences that may influence results. [Page 964]As such, individual differences that influence results are accounted for and the intervention can be modified or publicized with those characteristics in mind.

There are several types of SCED including multiple baseline, alternating treatment, and phase change designs. This entry focuses on a multiple baseline design, in which an intervention is staggered across behaviors, participants, or settings to demonstrate change occurs when and only when the intervention is introduced. Herein is the power of this design; it allows for strong causal inferences about the effect of the intervention on a designated outcome.

Multiple Baseline Design

Multiple baseline across behaviors is typically conducted within single subjects. In this design, an intervention is applied systematically across target behaviors (e.g., exploring whether a treatment increases the frequency with which a child ingests fluids, soft foods, hard foods, and chewy foods; see Nock, 2002) with each behavior serving as a control until it is intervened upon. In a between-subjects multiple baseline design, time serves as the baseline and participants are randomized to receive the intervention after differing lengths of time (e.g., a study exploring whether monetary incentives increase medication adherence for 10 adults with HIV/AIDS in which participants are randomized to waiting 2 or 4 weeks before starting to receive incentives). Whereas within-subjects designs assess multiple behavioral outcomes, between-subjects studies primarily focus on a single outcome (e.g., medication adherence). Finally, multiple baseline designs across settings may be conducted within- or between-subjects. In these designs, treatment is applied across independent situations and situations in which the intervention has not yet been introduced serve as the baseline comparator (e.g., evaluating whether an intervention reduces temper tantrums at home, at school, and in public).

...