This study compared three instructional approaches-- computer assisted instruction, cooperative learning, and teacher directed instruction--to determine their relative effectiveness in improving math performance of low achieving students. Additional information was collected on student time on task behavior to determine the relative impact of these treatments on this variable. An experimental research design was used. Ninety-nine rising sixth grade students were randomly assigned to one of the three instructional delivery groups for a five week summer remediation program. Classroom teachers self selected the treatment approach they used based on interest and personal experience. Additional training in the use of these strategies was provided prior to the beginning of summer school. Fourth grade students' scores on the math subtest Iowa Test of Basic Skills (ITBS) served as the baseline data for assigning students to one of the treatment groups. A subsequent ITBS math score was obtained on the same students as fifth graders with the latter score serving as the pretest measure. At the end of the summer program the ITBS math subtest was readministered to students to obtain posttest dependent measures on math concepts, math problems, math computations and math total. These data were analyzed with an ANCOVA with the fifth grade ITBS math total score serving as the covariate. While substantial academic growth was reported for all groups on the math total measure, it was found that no significant difference existed between the three groups on improving student performance on math concepts, math problems, math computations, or math total. On the time on task measure, students’ off task behavior observed was minimal and differences reported were not found to be significant. / Ed. D.