This study compares a self-assessment instrument and achievement test containing the same questions and administered at different levels of proficiency in order to investigate the reliability of second language self-assessment for student placement, diagnostic purposes, and achievement testing. Subjects were English as a Foreign Language students at Hebrew University in Jerusalem and students preparing to be English teachers at Teachers' Seminary. The instruments assessed vocabulary, grammar, listening comprehension, speaking ability, reading comprehension, and writing skills. Results suggest that (1) this form of test is unsatisfactory for foreign language placement of new students; (2) the most reliable self-assessment is found among students who had been studying for some time and actively use their skills; (3) self-assessment has great diagnostic value and can be used effectively as a formative test, during a language course; (4) students exposed to self-assessment become more conscious of their language strategies and thus better at self-assessment; and (5) this test type could be a valuable educational and psychological adjunct to teaching but is not yet proven reliable for second language achievement testing. (MSE)