The difficulty of developing and evaluating effective treatments in psychiatry and clinical psychology points out the inadequacy of current research methodology involving comparisons of large groups. An alternative approach firmly founded in the scientific method, but particularly appropriate to the study of complex behavior disorders, is the single-case experimental design. In this paper, examples of different single-case designs actually employed in applied clinical research are presented and discussed. Practical problems arising during the course of research are highlighted and some basic procedures outlined. General questions on variability, representativeness of finidings, and clinical versus statistical significance are briefly discussed.