摘要：The data from both single-case and case-series studies can be arranged in a matrix of N subjects by M tests. In theprototypical case study, we have a single subject who is given many neuropsychological tests, thus providing athorough examination of the subject’s abilities. The data form a 1 by (large)M matrix, and hence we can call this a“ long thin horizontal” study. Since in practice all single-case studies include data from unimpaired controls, theprototype is better described as a 2 by (large)M, matrix with control means forming an additional row. The typicalcase series, in contrast, is a “ long thin vertical” study, with a (large)N by (small)M matrix— many participants, but nota lot of tests. A case series with a single dependent variable is not terribly interesting (e.g. a collection of patientscores on a single verbal STM test), and so the viable minimum would be (large)N by 2, just enough to examine thecovariation between the tests. Of course, one can have studies that are not so thin, by testing several subjects onseveral tests. In this respect, the distinction between case studies with small numbers of subjects but several tests, andcase series with many subjects and fewer tests is in fact a continuum.
