Abstract
A case is made for repeated‐measures experiments, and particularly single‐subject designs, over randomized group studies. The former are argued to be more powerful than the latter, in that they take advantage of psychological “context” or “contrast” effects. Moreover, they provide functional relationships that best answer the kinds of research questions in which educational researchers are interested. Finally, they require the development of experimental, rather than statistical, control — an advantage that should help the field advance.