Hypothesis Testing in Psychoanalysis

Authors who undertake to assess the accomplishments of research in psychoanalysis (e. g., Wallerstein and Sampson, 1971) typically conclude that psychoanalysts have done well in generrating significant hypotheses relating to the theory and practice of their discipline but have failed dismally in testing these hypotheses (investigating them in controlled studies). Scientists who are prone to recognize only the latter kind of activity as genuine research conclude that serious efforts to substantiate the premises of psychoanalysis have yet to be made. Some of them question whether the concepts of the discipline are well enough defined to permit meaningful investigation at all and are doubtful whether psychoanalysis will ever be regarded as a legitimate science.

At the opposite pole are many practicing analysts who regard the generation of hypotheses from clinical observations as the most worthwhile kind of research and look on controlled investigations in psychoanalysis as either unproductive or, in principle, impossible for so complex a subject as the human psyche (see, for example, Shapiro, earlier in this journal). Clinically oriented professionals are typically not motivated to demonstrate