Criterion-referenced and norm-referenced tests are compared, and norm-referenced tests are found to have inadequacies as a basis for evaluation or instructional improvement. Disadvantages of norm-referenced tests, which measure a student's performance relative to the performance of other students, are: (1) mismatches between local curriculum and what is measured by standardized tests; (2) anxiety and frustration caused by imprecise knowledge of needed skills improvement; and (3) tendency to omit test items which measure the knowledge teachers consider most important, in order to obtain a large degree of response variance. Advantages of well-constructed criterion-referenced tests include pinpointing necessary curriculum changes; devising effective instructional sequences to attain needed competencies through task analysis, task description, and practice; and aiding in course evaluation. Criterion-referenced tests may be obtained from one of several sources: commercial test publishers, local school district production, or test development agencies. Attributes of well-constructed criterion-referenced tests include a description of what the measured performance means; a sufficient number of test items to measure each specified behavior, and reliability and validity. (MH)