Determining whether outcome scores are close to normative means is a commendable way to assess clinical significance. However, to be valid, norms require samples that are carefully chosen to be representative of relevant populations. The Eyberg Child Behavior Inventory norms used by Sheldrick et al. were not representative of relevant populations, thereby casting doubt on conclusions concerning the clinical significance of outcomes. In addition to being used for categorical judgments of whether outcomes approximate a normative mean, valid norms are needed for deciding who needs treatment; for evaluating outcomes for problems too severe to approach the normative mean; for making diagnostic criteria more sensitive to age, gender, and other differences; and for evaluating aspects of functioning beside those targeted for treatment.