We present results from a randomized study of a well-defined use of computers in schools: a popular instructional computer program for pre-algebra and algebra. We assess the program using a test designed to target pre-algebra and algebra skills. Students randomly assigned to computer-aided instruction score 0.17 of a standard deviation higher on pre-algebra/algebra tests than students randomly assigned to traditional instruction. We hypothesize that the effectiveness arises from increased individualized instruction as the effects appear larger for students in larger classes and in classes with high student absentee rates.