A complex problem exists about how to promote the best interests of children as a group through research while protecting the rights and welfare of individual research subjects. The Nuremberg Code forbids studies without consent, eliminating most children as subjects, and the Declaration of Helsinki disallows non-therapeutic research on non-consenting subjects. Both codes are unreasonably restrictive. Another approach is represented by the Council for the International Organizations of Medical Science, the U.S. Federal Research Guidelines, and many other national policies. They allow research ethics committees or institutional review boards to authorize studies with acceptable balances of likely benefits and harms, but neither clarify how to balance them nor explain the meaning of pivotal concepts, like "minimal risk." Paths to the improvement of balancing or consequentialist approaches include (1) improving standardizing of risk assessment, (2) rejecting crude utilitarianism, (3) identifying and justifying normative or moral judgments, and (4) acknowledging extra-regulatory thresholds and deontological or non-negotiable duties to children.