This paper reviews the problems of identifying the kinds of risks to which subjects in research projects are exposed, establishes the role of federal funding agencies in risk identification and suggests ways to protect children who are research subjects. The risks cited are primarily psychological, sociological, physiological, and/or academic. Examples of risks are overtesting of children, creating dependency, and labelling. A process for protecting human subjects from harm in research should be established at each institution receiving research funds. The researchers, possibly aided by a local committee, should be responsible for seeing that children who are research subjects are not at risk. The protection of children as research subjects is a complex effort requiring the participation of many groups and organizations at the federal, institutional, association and research project levels. (CM)