A discussion of student and program assessment in the context of Chinese language instruction looks at theory and methods of assessment, and proposes that norm-referenced language proficiency testing is inadequate to evaluate student achievement and program effectiveness. It is further proposed that criterion-referenced tests are more appropriate in assessing both student learning and program effectiveness in Chinese language programs. Deficits identified in norm-referenced tests currently in common use include discrepancies between instructional objectives and test content, descriptive ambiguity, inadequate test interpretation, threat to content validity due to built-in item selection that systematically eliminates "easy" items, and inappropriateness for summative selection. It is argued that criterion-referenced tests, by comparison, are based on instructional objectives, have descriptive clarity, can be interpreted according to degrees of mastery, and can include some easier items because there is no intent to maximize differences among individuals. (Contains 19 references.) (MSE)