Stanford Achievement Tests (Science) were examined to point out limitations of standardized achievement tests. Several factors teachers need to consider when using standardized tests were identified. Tests can be effective only if they measure the same content areas that make up the science curriculum. Good reading comprehension is needed for students to score well on science achievement tests (SATs). Finding SATs measuring the same topics as the school science program may be difficult. A well-designed, teacher-prepared test may be better than SATs which do not fit the curriculum. A series of tests (including those of former years) may be desirable. Tests from whole classes may be effective in evaluating a school science program. Scores may be improved by guessing and not leaving items blank. How a student feels on a day may affect test scores. Teaching to test or making students test-wise may raise school achievement scores but does not reflect the true condition of the science program. SATs are statistical measures based on probability and with enough data reliable inferences can be made. Overdependence on SATs can lead to designing science programs to help students score higher on the test. Tests should help teachers measure student outcome and not become curriculum determiners. (Author/JN)