Magnet schools have become increasingly popular as alternatives to traditional curriculum offerings and as tools to promote voluntary desegregation in urban school districts. Most evaluations of magnet schools have been concerned with monitoring objectives and compliance with regulations. Few reports assess the impact of magnet curricula on student achievement. The evaluation of achievement in magnet programs cannot be interpreted when there are no comparisons of magnet students with nonmagnet students or with scores district-wide. The Austin (Texas) Independent School District implemented the Science Academy of Austin, a high school science, mathematics, and computer technology magnet program in 1985. At the end of the first year the tenth-grade magnet students' achievement gains did not significantly out-gain their district-wide counterparts in science. Two reasons were hypothesized: (1) the Tests of Achievement and Proficiency (TAP) Science test did not address the curriculum taught to tenth-grade magnet students; and (2) the TAP Science test was not measuring what the test publishers said the test should measure. A factor analysis of the TAP Science test gave support to both hypotheses. Evaluators should examine their implicit assumptions about the appropriateness of using results of standardized tests to evaluate special programs. (BAE)