This paper reviews several issues and problems associated with the creation and application of tests and decision models for determining entry/exit, eligibility, placement, treatment, and reclassification procedures used to remedy the limited English proficiency of students from homes where English is not the primary language. Two concepts critical to the assessment process are discussed: language dominance and language proficiency. It is argued that the concept of language proficiency is not only more linguistically and scientifically sound, but also more amenable to mathematical or statistical manipulation because of the known properties of the test score distributions. A number of the ideas used in this argument are used to review some of the problems with current testing practices related to eligibility, placement, and reclassification. It is suggested that the failure to work from a common set of definitions and principles has compromised not only the process of entry/exit but also both the evaluation of Title VII programs and research on the effects of bilingualism. The moral issue of assessing students is discussed. An operational definition of limited English proficiency is given that includes the following components: limited English proficiency student, comparable students, language proficiency, and probability of success. Contains approximately 75 references. (LB)