Single-subject designs are a research method that test for an independent variable in a small number of subjects, typically one. The method is thought to have originated from behavioral psychology and is preferred for experiments involving behavioral change or modification because the outcomes are tested periodically from the beginning of the experiment to the end. Single-subject designs include at least one phase where there is no treatment and one phase where there is treatment. The test subjects, therefore, act as their own control. Single-subject designs are relevant where behavioral change needs to be observed, as one can not only measure the change, but also the degree of change. They also include a number of draw-backs and limit, such as variable baselines and permanent changes to the baselines. Single subject designs include AB designs, multiple baseline designs and withdrawal designs.
Withdrawal designs (also called reversal designs) are an extension of the AB designs. The AB designs contain only one baseline and one treatment and are the simplest of the single-subject designs. Withdrawal designs contain more than one baseline (A) and more than one treatment phase (B). Examples of withdrawal designs include ABAB, ABABAB, and BAB. The ABAB design, for example, tests for a baseline (A), then a treatment (B), then a withdrawal (A), and finally a re-introduction of treatment (B). If one was to determine the effect female pheromones on a male dog’s heartbeat using the ABAB design, then the research question would be: Does the presence of female pheromones have an effect on a young male dog’s heartbeat? Does it rise or fall?
The hypothesis would be “The presence of female pheromones has an effect on a male dog’s heartbeat, causes the heartbeat of a young male dog to rise.” For the hypothetical experiment, the puppies will be between three and four months old. The experiment would be thus: the puppies kept at conditions that stabilize their heart rates (no distress, loud noises, arousing smells, etc.) The room in which each dog is kept is cleaned, and air filtered to prevent external, unintended scents from corrupting the results of the experiment. At this stage, when the heart rate is stable, the first measure is taken (A). After ten minutes are over, the pheromones are introduced into the air. The puppy’s heart rate is again measured for ten minutes (B). It is left in the room for a period of ten minutes, and the heart rate is observed the whole time. Next, the female pheromones are again cleared from the room, perhaps by a ventilation system. After the air has been tested for the pheromone and judged to be clean, the puppy’s heartbeat is again tested for the same period (A2). After ten minutes the pheromones are again released into the air, and the heart rate observed for ten minutes (B2).
The heart rate at A is the baseline. If the heart rate rises to a particular level at B, falls to the same level as A at a2 and then rises to the same level to the same level as B at B2, then one could conclude that female pheromones do indeed have an effect on male dogs before the onset of sexual maturity. The above hypothetical study is not appropriate in the ABA study, because one would have no way of confirming that indeed the pheromones caused the heart rate to rise at B. ABAB design therefore offers more control that ABA design.
Adding more A and B phases isn’t always possible or even ethical. Therefore for some studies one might settle for a simpler design such as ABA than a more complex one such as ABAB. While ABA could be used in situations where a mental or health disorder or undesirable condition is successfully eliminated, ABAB could not, as that would mean re-introducing the successfully treated disorder. Hence in such a situation ABAB is limited by ethical standards while ABA is not.
In another hypothetical study, the effects of reinforcement learning on the academic performance of a child are tested and measured. Using the ABA design, the baseline is established with the child’s number of successfully answered questions in an appropriate arithmetic test (A). The child receives praise or a form of present for correctly answered questions. The next test the child does (perhaps homework) is checked for correctly answered questions. The child does not receive praise for these (B). The next time he does homework; it is checked for correctly answered questions. At this point, he receives praise and presents again, and the homework is again checked for correctly answered questions. Here the ABA design is appropriate while all the rest are not, because the child might simply get better at arithmetic and if he does it a lot (ABABAB). In this case the baseline is also permanently affected and may not revert to its original position, simply because the child improves at doing arithmetic every time he puts some effort. Irreversibility of the baseline affects all withdrawal designs. A threat to the internal validity of this study include: the child could have an external motivation for performing well in arithmetic. One cannot guarantee that the child will not be affected by external factors. Although there are no ethical issues associated with this study, one socio-cultural issue arises: children, being differently socialized and brought up, might react differently to the study. Hence it would not be sufficient to carry out the test on only one subject. The credibility and validity of the test is increased with diversity.
Works Cited
Richards, S. B., Ronald, T., & Rangasam, R. (2013). Single subject research: Applications in educational and clinical settings. Boston: Cengage Learning.
