The paper argues that no criterion‐referenced assessment system can achieve both reliability and validity at one and the same time. It shows that the reasons for this are conceptual, and hence that empirical research into the‘problem’ is a waste of money and effort. Considerable discussion is devoted to ideas of knowledge and understanding, and to proper educational objectives pertaining to these. Much reference is made to the current National Curriculum context in the United Kingdom, and conclusions are drawn for appropriate educational policy in this