Single-Subject Designs

Single-subject designs involve the intensive study of individual subjects under various conditions, in various environments, or some combination thereof. The subject of interest is often a single organism, [Page 904]such as a client in a clinical setting, but could also be a group of individuals acting as a single unit, such as a department within an organization or a school within a district. This focus on the individual is reflected in the alternative labels for single-subject designs, including single-case designs, N of 1 designs, and within-subject designs. Although single-subject designs focus on the individual, studies using these designs typically include several participants.

Single-subject designs can be used in correlational or experimental studies, but we will focus on experimental designs. These designs share some basic features with group (or between-subject) experimental designs: An independent variable (IV) is manipulated, a dependent variable (DV) is measured, and nuisance variables (NVs) are controlled. Unlike participants in group designs, participants in single-subject designs receive all levels of the IV, while the DV is measured repeatedly under those levels. The values of the DV under each level of the IV (e.g., a control condition and one or more treatment conditions, or interventions) are compared, and causal inferences regarding the effects of the intervention(s) are then made. In these designs, each participant serves as its own control.

Single-subject designs provide useful and flexible methodological options when (a) one cannot obtain appropriate numbers of participants for group designs (e.g., the population of interest is relatively small, such as persons with a rare medical condition), (b) one seeks to assess treatment effectiveness in particular individuals (e.g., the progress of specific clients is of interest), or (c) one cannot ethically use group designs (e.g., withholding treatment from individuals in a control group is not an option). Practical reasons may also necessitate the use of single-subject designs.

The specific details of single-subject designs vary, but most such designs share two common features. First, the DV is measured repeatedly across time; how often it is measured depends on many factors, including available resources and the nature of the study. Repeated measurement permits the researcher to track changes in the DV and, if necessary, to modify treatment conditions if the DV does not change in the desired direction. Second, the level of the DV in the absence of treatment is established during a baseline (or A) phase, which typically occurs at the beginning of the study, before treatment is administered, although there are some exceptions.

The baseline phase serves three important functions: (a) It provides information regarding the pre-treatment level of the DV; (b) it provides information on the predicted level of the DV if treatment is not applied; and (c) it provides control data with which treatment data will be compared, with differences in the DV between baseline phase and the treatment (or B) phase providing evidence for the effects of the intervention. Figure 1 presents fictitious data to illustrate these functions. The DV is the number of accidents per week at a job site, and the intervention is a program that provides workers with incentives for every day without an accident. Panel A shows that the baseline level of the DV hovers around 15 accidents per week. From these data, we can predict that the number of accidents would probably remain the same if the intervention were not introduced. Panel B depicts a decrease in the number of accidents on introduction of the intervention. The change in the DV from the baseline phase to the treatment phase provides evidence, albeit limited, of the intervention's effectiveness.

...