A review of literature serves as the basis for a discussion of various aspects of criterion-referenced tests. The aspects discussed are: teaching and evaluation objectives, criterion- and norm-referenced measurement, stages in construction of criterion-referenced tests, construction and selection of items, test validity, and test reliability. Examination of the literature shows that criterion-referenced and norm-referenced measurement share some features, including dogmatism in its introductory stage, but that criterion-referenced measurement (CRM) has the advantage of providing an exact description of an individual's performance level in an entire domain, not only on the presented items. When test construction and interpretation are done correctly, the desired result occurs because of CRM's emphasis on the conceptualization of measured domains. Because tests are the most concrete way of signaling to teachers and students what the desirable content and forms of teaching are, and because testing has a limited scope, it is suggested that focusing on testing in instructional development may be more effective than focusing on curriculums and teaching materials. (MSE)