SINGLE SUBJECT DESIGN
Single subject design and analysis is an experimental ideal. An individual is studied under a number of conditions, and the analysis is performed on the data of this individual. The main advantage is substantive: maximal congruence with psychological phenomena. A collateral advantage is that longer-term investigations may be practicable, unfolding phenomena barely present in the common one-session experiment. There is also the statistical advantage that error variability will be even less than with repeated measures design.
Single subject design has always been a mainstay in perception. One reason is that many perceptual phenomena can be embodied in stable state tasks, so that responses under one condition are not confounded with order effects from previous conditions. One subject can thus provide a complete data pattern across all experimental conditions. Many studies use just two or three subjects, with results presented separately for each. Many results can be generalized on the basis of extrastatistical, background knowledge about similarity of sensory–perceptual process across individuals.
Single subject design has also been useful in diverse other areas. Among these are classical and operant conditioning, judgment–decision theory, physiological psychology, behavior modification, and medical science. Also notable are studies of unusual individuals, such as language disorders caused by brain damage that shed light on the multiple functional components of language.
A pall hangs over single subject design and analysis. This topic goes virtually unmentioned in current graduate statistics texts. Whole areas of experimental analysis that could benefit from this approach make little use of it. You can check yourself how very few single subject studies appear in any issue of any journal published by the American Psychological Association. a
-307-
