This study analyzes the Arizona policy of utilizing a single assessment of English proficiency to determine if students should be exited from the ELL program, which is ostensibly designed to make it possible for them to succeed in the mainstream classroom without any further language support. The study examines the predictive validity of this assessment instrument on ELLs' performance on state required academic achievement tests at three grade levels. It finds that at subsequent grade levels after redesignation, the "one-test" AZELLA becomes less predictive of academic achievement, That is, the test over predicts student achievement, suggesting that many students may be under-served due to their scores the test. This finding calls into question Arizona's "one-test" procedure for redesignating ELLs to a non-service category. Given the large and increasing size of the ELL student population in Arizona, the current focus on testing and accountability, and the documented problems in current assessment practices, improvement in instruments and procedures is critical. These improvements are necessary at all phases of the assessment process, but as this study indicates, the present policy is likely denying services these students need and violating the rights of these students to an equal educational opportunity. (Contains 3 tables.)