Empirical norms are a traditional and accepted tool for reporting performance. Other methods (e.g., statistical regression) are used more often in a research context to make conditional statements about performance given a certain demographic context. This paper illustrates a method for the construction of differentiated school norms and the computation of percentiles on the basis of demographic background factors. These differentiated norms facilitate reports of performance that take into account demographic conditions that are beyond the control of schools. With this approach each school is located at the median of its own norm group with regard to an index of socioeconomic status. The computation of percentiles is illustrated with science achievement test scores for all eighth-grade students in 1,573 public schools in California. The effects of varying the size of the norm group are illustrated with reading achievement test scores for all 12th-grade students in 797 public high schools in California. The computing and reporting of differentiated norms are complex and may be difficult to communicate to general audiences. Local authorities in California, however, have reported few difficulties so far in communicating the concept of differentiated norms. Two tables and one figure illustrate the study data. (Author/SLD)