For many years now I have been required to give a series of elementary lectures on statistics to medical students about to undertake a postgraduate course in psychiatry. The declared aim of the course, for which very limited time was available, was to provide the students with some initial understanding of the statistical terminology and elementary techniques to which other teachers, in particular psychologists and sociologists, would be likely to refer in the course of their lectures. The task was tricky for two reasons. In the first place most of the students involved, despite their best...