The requirement that a subject be competent as a condition of valid consent to participate in research has been accepted by most students of legal and ethical problems of human experimentation. "Competency," however, has lacked a clear and generally agreed on standard. There are four commonly used standards for competency: evidencing a choice in regard to research participation, factual understanding of the issues, rational manipulation of information, and appreciation of the nature of the situation. These standards can be arranged hierarchically such that each represents a stricter test of competency. The decision as to how rigorous a standard for competency is desirable cannot be made on psychiatric grounds. It requires consideration of the policy goals on hopes to attain. Empirical research helps demonstrate the consequences of choosing a particular standard but cannot replace the need for achieving consensus on policy goals.