STATISTICS
R. J. Prescott and
Student
. . .Perhaps the medical profession should take a close look at its attitude towards statistics. The medical statisticians are thin on the ground, and it would be unrealistic to expect all papers with some statistical content to be refereed by them, even if this was thought desirable. Another way to tackle the problem is through undergraduate and postgraduate medical education. Since most medical students will not become research-workers, it is debatable whether the computational complexities of a test form a relevant part of their medical education. However, all doctors should keep up to date with developments reported in the journals, and to do so adequately they must learn to judge published work critically. It is this aspect of statistics which appears to be in most need of emphasis in undergraduate medical education.
For the aspiring researcher, the most important statistical skills to be learnt at postgraduate level are the design of experiments, surveys, and clinical trials. Without them the techniques of statistical analysis, which also need to be learnt at this time, may have poor material to work on.
Unless moves of this sort occur the present situation is likely to continue. Many research-workers will spend much of their time on work which is doomed in advance to be of no medical or scientific merit; valuable research facilities will be inefficiently used; the literature will continue to produce articles where the conclusions do not follow from the data. Worst of all, such articles will continue to be believed.
