The costs, effects, and utility of mathematics instruction supplemented by microcomputer presentations were compared with those of traditional classroom approaches for grades 3 and 5. The study was completed at a single school in a rural area of Saskatchewan, Canada, where two classrooms at each grade level were used in a pretest-posttest design. In both grades, students who received microcomputer instruction scored significantly higher on standard tests of total mathematics achievement and computer literacy than did those who received traditional instruction. Students in all groups scored about the same on a survey of mathematics attitudes. Based on utility ratings obtained from school board members, microcomputer assisted instruction was found to have greater cost utility in both grades. A strong case is suggested for the cost-effectiveness of pairing students on microcomputers placed in classrooms and used to provide adjunct instruction. (8 references) (Author/DB)