This paper reports the findings of a study that assessed the effect of two classwide instructional grouping adaptation strategies on the math performance of English-language Latino learners in an urban public school in the Southwestern United States. An ABAB within-case design was implemented over a 12-week period to contrast the effectiveness of these adaptation strategies on the math performance of 19 low-, average-, and high-achieving first grade Latino students (ages 6-8). All instruction was conducted in Spanish, with the exception of 35 minutes in which students received English-as-a-Second-Language instruction. In Phase A, students worked independently in small heterogeneous groups of 4-5 high-, average-, and low-achieving students. In Phase B, peer tutoring was used following a teacher-led instruction. Dyads included students with different performance levels; students in each dyad alternated roles as tutors and tutees. Student math achievement was assessed for the whole group as well as for distinct performance level subgroups. The results indicated the students' math performance was significantly higher during the peer tutoring phases than when in heterogeneous small groups. (CR)