Sir

The United Kingdom's experimental programme that teaches students research skills is a step in the right direction1. The experience may also serve as a model for the United States. Many American graduate schools seem to teach students skills specific to their fields of specialization, but not general principles of research methodology. It is time they considered adding such courses to their graduate programmes.

Scientific research consists of sophisticated intellectual activities that are poorly understood. Rational or logical approaches often fail, whereas irrational methods are sometimes productive. Many scientific discoveries are made by chance rather than as a result of planned research. It is frequently intuition, not rational or conscious thinking, that leads to an innovative theory. This is why many consider the methodology of research an art that cannot be taught. For the same reason, it might be difficult to come up with a method, as desired by Boryeu Mao2, that measures fruitless research activities against those that are productive.

Nevertheless, I believe that some general principles of research methodology do exist, and an introduction to such basic knowledge (logic, philosophy of science, skills of thinking and so on) can help young scientists to improve the efficiency of their research. These methods and skills are universal and can be applied to any field. For instance, the reductionist programme has been widely used in research, even though some may not know the term reductionism. It is therefore useful for researchers to understand that this approach is not appropriate for all problems in science3,4. In his time-honoured monograph5, W. I. B. Beveridge offered a vivid account of methodology for making scientific discoveries from a psychological perspective. I recommend this work to science students.