The psychoanalyst who wishes to do research will find little encouragement in the ordinary book on methods of research. Design of research all too often is treated as synonymous with statistical design and obviously is not suited to the sorts of data the analyst encounters or to the single, or at most few, cases that he treats. The present volume is written in a different tradition and clinicians will find congenial the points of view it advances about studying individual subjects intensively, about the disadvantages of pooling data from many subjects, and about the statistical approach to purging data of disturbing variables in order to produce a kind of purified, idealized situation which incidentally is not true to nature. Sidman correctly notes that rather than being a nuisance to be wished away or balanced out, variability can reveal the lawfulness of behavior when adequate experimental control has been achieved. He offers no list of procedural rules for research but rather views the doing of research as a highly personal, even artistic, matter for which one