Duncan Neuhauser, Mireya Diaz and Iain Chalmers (JRSM 2008;101:381–383) are puzzled that one of the great pioneers of clinical trials, Russell LaFayette Cecil, failed to include a chapter on trial methodology when he went on to edit his best-selling Textbook of Medicine – but perhaps they are being diplomatic.1 If Cecil thought most doctors did not need to worry their heads about science he was merely reflecting the profession's long-standing ambivalence to science after two millennia of reliance on Galenic teaching and personal experience. The fact that there are still qualified physicians who endorse Prince Charles' approach to medicine (embrace science when it suits but not when it doesn't) speaks volumes. As Lord Darzi points out in the same edition (JRSM 2008;101:342–344), the term 'evidence-based medicine' did not enter the medical literature until the 1990s; and there is still a lack of knowledge about fair tests of safety and efficacy2 – as I can testify from trying to hand out charitable money to would-be medical researchers, a disconcerting proportion of whom have a poor grasp of design, analysis and statistical power. Things may not be much better in future. Some undergraduates seem to be taught more sociology than methodology; surely a gap in medical education which suggests that the GMC and several deaneries concur with Russell LaFayette Cecil's omissions of generations past.