Statistics, Science, and Psychiatry

I. Introduction  During the past 20 years clinical trials have come into prominence and fashion both in general medicine and in psychiatry. In the latter, trials of tranquilizing drugs have stimulated the application of statistical techniques, and today most journals contain more than a fair share of mathematical analysis and evaluation. Indeed, one can wonder whether such an evaluation has not become the fig leaf of scientific respectability without which a relevant conclusion cannot be drawn. It is proposed in this paper to draw attention to some of the limitations of statistical evaluations as they appear in the literature, to examine the relation between science and statistics, and to examine the contribution of statistics in the field of psychiatry.

II. Limitations  The conclusions of clinical trials are usually presented as follows: “There is (or is not), a significant difference between treatments A and B for