Single-subject research

Single-subject research is a group of research methods that are used extensively in the experimental analysis of behavior and applied behavior analysis with both human and non-human participants. Principal methods in this type of research are: A-B-A-B designs, Multi-element designs, Multiple Baseline designs, Repeated acquisition designs, Brief experimental designs and Combined designs.

These methods form the heart of the data collection and analytic code of behavior analysis. Behavior analysis is data driven, inductive, and disinclined to hypothetico-deductive methods. Statistical methods have been largely ignored.

Experimental questions are decisive in determining the nature of the experimental design to be selected. There are four basic types of experimental questions: demonstration, comparison, parametric, and component. A demonstration is "Does A cause or influence B?". A comparison is "Does A1 or A2 cause or influence B more?". A parametric question is "How much of A will cause how much change or influence on B?". A component question is "Which part of A{1,2,3} - A1 or A2 or A3... - causes or influences B?" where A is composed of parts which can be separated and tested.

The A-B-A-B design is useful for demonstration questions.

An AB design is a two part or phase design composed of a baseline ("A" phase) with no changes, and a treatment or intervention ("B") phase. If there is a change then the treatment may be said to have had an effect. However, it is subject to many possible competing hypotheses, making strong conclusions difficult. Variants on the AB design introduce ways to control for the competing hypotheses to allow for stronger conclusions.

The reversal design is the most powerful of the single-subject research designs showing a strong reversal from baseline ("A") to treatment ("B") and back again. If the variable returns to baseline measure without a treatment then resumes its effects when reapplied, the researcher can have greater confidence in the efficacy of that treatment. However, many interventions cannot be reversed, some for ethical reasons (e.g., involving self-injurious behavior, smoking) and some for practical reasons (they cannot be unlearned, like a skill).

Further ethics notes: It may be unethical to end an experiment on a baseline measure if the treatment is self-sustaining and highly beneficial and/or related to health. Control condition participants may also deserve the benefits of research once all data has been collected. It is a researcher's ethical duty to maximize benefits and to ensure that all participants have access to those benefits when possible.

The A-B-C design is a variant that allows for the extension of research questions around component, parametric and comparative questions.

...
Wikipedia