This summative evaluation of magnet programs employed a quasi-experimental design to investigate whether or not students enrolled in magnet programs gained any achievement advantage over students who were not enrolled in a magnet program. Researchers used Zero-One Linear Programming to draw multiple sets of matched samples from the non-magnet student population to serve as multiple control groups for the research effort. Whenever a matched sample was generated, the analysis of covariance was subsequently used to control for the effects of the possible pretest score difference between the two groups on the outcome variable (posttest score on reading or mathematics). The mean of the effect size values, across 200 matched-sample analyses, was used to investigate the effects of the magnet program treatment on the reading and mathematics performance of magnet students. The results for the elementary school magnet programs showed that when students' demographics and initial abilities were accounted for, only the French immersion program had a likely positive impact on students' reading or mathematics performance. Nevertheless, the rest of the magnet programs had minimal, if any, positive effect to promote higher reading and mathematics test scores of magnet students. Appended are: (1) Assumptions Underlying the Use of ANCOVA; (2) The Meaning of the Effect Size; and (3) technical details. (Contains 2 tables and 15 figures.)