Single-case experimental designs

The central aim of science is causal explanation. In meditation research, we want, for instance, to find out about causal relationships such as “If I meditate, my worries diminish," or “If I practice mindfulness meditation, my mindfulness will grow" The basic procedure for arriving at causal explanations in science is to perform “true” experiments, in which one or more independent variables are manipulated (e.g., meditation training: Yes or no) and all other potential causal factors (such as gender, age, education, motivation, etc.) are controlled for. In true experiments this control is achieved by dividing participants randomly between the experimental group(s) (e.g., meditators) and a control group (e.g., non-meditators). This randomization procedure guarantees that all variables that also might have an impact on the dependent variable(s) on average have the same or similar effects in both the experimental and the control group and so cannot systematically influence the difference in outcomes for the two groups. If, then, a difference between groups (e.g., meditators and non-meditators) is found, it can be concluded with high confidence (the height of this confidence depends on how well all parts of the study were operationalized) that it must be due to the manipulation of the independent variable (e.g., some people meditated and others did not). Randomization works well with large samples, less well with small samples, and not at all well with individuals.

The difference between any single-case design and single-case experimental designs consists in some kind of randomization that is a central ingredient in the latter (for a good introduction and overview see Barlow et al. 2009). This randomization obviously cannot be done over people. Instead it is done over time for a given person. The basic idea in single-case experimental designs is that the treatment(s) or intervention(s) are administered at randomly chosen points or intervals in time and then compared to baseline intervals or intervals of other treatments. Similar to randomization over people in group designs, randomization over time in single-case experimental designs controls for causal influences other than the independent variable(s) in question.

Probably the best known kinds of single-case designs are A-B-A designs or variations thereof (e.g., A-B-A-B, etc.) that begin with a baseline (A), introduce a treatment (B), and then withdraw the treatment again (A). Such designs are appropriate for examining whether a treatment is effective in principle but they target treatments that are not assumed to have lasting effects, as would be expected for meditation practice. However, a variation of this kind of design might be of some value in meditation research and will be briefly discussed later. More appropriate for examining the effects of meditation are multiple- baseline designs and, for special kinds of research questions, alternating- treatment designs.

< Prev   CONTENTS   Source   Next >