Published online by Cambridge University Press: 06 October 2014
Proponents of between-group experimentation are quick to point out that parametric methods are superior to the experimental analysis of individual responses on the grounds that group statistical designs have external validity by virtue of random sampling and standardised rules and conventions for hypothesis testing. In contrast, advocates of single-case experimental designs invariably invoke the argument that a major limitation of between-group methodology is that the averaging of results masks individual differences. The data presented illustrate some strengths and weaknesses of both procedures and demonstrate that generalising from groups to individuals can be practically useless.