Evaluation of bilingual education programs is limited by the absence of appropriate standards against which program success may be gauged. Normative and criterion models have been used, inadequately. The current paper revisits the normative model, and examines the relative advantages of using norms based on the performances of monolingual-curriculum students or program participants. Data from New York City's standardization administration of the Language Assessment Battery (LAB) were analyzed. The results show that use of monolingual-curriculum norms advantage program students in listening skills, and also advantage students in higher grades. Elsewhere, program participant norms afford greater advantage to the program students. These findings are consonant with language acquisition theory (Cummins, 1979). (Author)