Commentary: On the limited role of the ''single-subject'' design in psychology: Hypothesis generating but not testing

Journal/Book: J Behav Ther Exp Psychiat. 1999; 30: the Boulevard Langford Lane, Kidlington, Oxford Ox5 1GB, England. Pergamon-Elsevier Science Ltd. 21-22.

Abstract: The ''single-subject'' design (which really denotes a design that employs too few subjects to allow statistical inferences concerning significance to be made) is useful only for the generation, but not for the testing or evaluation, of hypotheses concerning any psychological function. Those hypotheses that may be suggested by the ''single-subject'' design include ones about within-subject developmental effects (which may need to be studied over multiple sessions), as well as individual-difference-related variables. To adequately test any of those hypotheses, however, it is necessary to employ designs that vary both within- and between-subject factors, and that also examine various correlations in,such a way that all effects (including correlational ones) can be evaluated by conventional modes of statistical inference.

Note: Editorial Furedy JJ, Univ Toronto, Dept Psychol, Toronto, ON M5S 3G3, CANADA

Zurück | Weiter