This article analyzes potential problems in using standardized achievement tests in early childhood education settings. The first problem discussed is whether the outcomes typically measured by standardized achievement tests are the outcomes about which early childhood educators are most concerned. Second, it is argued that a concern for test reliability must be balanced against whether the tests are appropriately specific or general and what kinds of skills they measure, etc. Third, the problem involved in measuring only intended program outcomes is noted. It is suggested that educators must be concerned that programs meet certain minimal standards (intended or not), that the programs optimize their intended goals, and that the tests should reflect these concerns. Fourth, more specific problems of testing are pointed out, i.e. (1) test items being more appropriate for some curricula than others, (2) items requiring more specialized knowledge than they were intended to test, (3) items being developmentally inappropriate, (4) preparation for the test taking time away from other valuable activities, and (5) the difficulty of testing for traits with delayed effects. Fifth, the difference between broad-domain and narrow-domain tests is explained, and the dangers of organizing the curriculum around these tests are noted. Suggestions are made on using tests and explaining the uses and limitations of tests to the public. (BH)