Before selecting an achievement test, it is essential to determine its purpose. There are eight purposes, relating to three educational decision making contexts: (1) instructional managerial (diagnosis, course placement, career guidance); (2) screening (selection and certification) and (3) programmatic (survey assessment, formative evaluation, summative evaluation). Test publishers and developers should report score consistency (reliability), validity, and logistics of administration. High reliability coefficients (.90-.99) are needed for screening; placement and guidance require a .80-.99 range; a lower range .70-.99 is acceptable for diagnosis and programmatic decisions. Teachers can determine validity by linking test content with their own predetermined instructional objectives, or with other tests and course grades. Publisher reported validity clusters around .90. Teachers and subject experts are the best validity judges of instructional and screening tests; administrators and teachers should judge validity for programmatic decisions. Scores can be referenced to a norm group or to instructional objectives. Logistics include thoroughness of instructions for administration, security, scoring, and costs. A list of 102 currently available tests, including achievement batteries, and tests of reading, mathematics, life skills, and language arts is appended. A publishers' directory and bibliography of test review sources are also appended. (CP))