Professionals who make decisions in child abuse cases often deny that their decision-making has a moral basis, claiming that they have a “non-judgmental” attitude, and that they only make decisions in the “interests” of their clients, reaching their conclusions on an assessment of the circumstances. This paper considers the implications of this for practice, and particularly for the ways in which cases are reported to the authorities. The study suggests that the general public are reluctant to report cases to the authorities, or even to suggest that professionals should intervene in an authoritarian way. This may arise from suspicion, often in effect justified, that the child will be removed and placed in an institution. Normally the decision to remove a child from home is taken in two stages: first professionals decide whether a child has been abused or whether he is in acute danger of being so; only if this is so can the second decision be taken, namely whether to remove the child or offer some less drastic form of help. This paper is primarily concerned with the first issue and with forms of abuse which are not in themselves life-threatening. Five types of criteria are considered for determining whether child abuse has taken place and the implications for each for public perception of professionals and for practice. The types are: legal, social norms, professional expertise, client subjectivity and no criteria. All present difficulties and may sometimes be in conflict with one another. The paper argues for standards to be clearer, criteria more available and for a shift of resources from therapy to education.