This study examined the effects of grouping students by ability or achievement on middle school students' academic achievement. Mathematics achievement scores from the California Test of Basic Skills (CTBS) were obtained from 25 seventh graders randomly selected from a group of 80 who had received instruction in an ability-grouped setting for mathematics instruction during third grade and in an inclusive setting during the sixth grade. The mathematics scores in the third grade ability-grouped setting were compared to those from the sixth grade inclusive setting. Also, a written survey was completed by kindergarten through eighth grade teachers in five elementary schools and one middle school. Mathematics achievement results indicated that there was no significant difference between math scores in the third grade ability-grouped setting and the sixth grade inclusive setting. Teacher survey results revealed that while most teachers have moved away from rigid forms of ability grouping, some still group for subjects such as reading and mathematics. Those who still group students by ability revealed that they continue this practice because they believe that delivery of instruction is easier when students are on the same instructional level. Students' previous grades, current test scores, and teacher perceptions were used to group students by ability. Teachers who had moved away from grouping indicated that the transition was made easier by employing alternative methods and styles of instruction--especially a multisensory approach--but also including cooperative learning and mastery learning. (Five appendices include t-test data and a copy of the survey. Contains 31 references.) (Author/KB)