Cooperative learning in reading was undertaken with 1299, Grade 6 students and 1275, Grade 8 students drawn from 120 classes in 10 high school settings. Grade 8 students acted as peer tutors to Grade 6 students in a 16-week long randomized controlled trial (60 intervention and 60 control classes). Data analyzed through hierarchical linear multi-level modeling indicated significant benefits of the technique on reading attainment to students in the bottom reading attainment decile at pre-test. These students recorded a significant Effect Size=+0.49, compared to students in the control group. Results indicated that cooperative learning might be effective at closing the attainment gap for low attaining students, but effects did not generalize in the manner previously reported in elementary school.