To investigate the effect of heterogeneous and homogeneous instructional groupings used in a middle school eighth grade on selected student behaviors, 120 students were assigned to a heterogeneous group and 120 students to a homogeneous group based on several factors. The heterogeneous group was comprised of all students enrolled in the school band, approximately 20 at-risk students, and other students of differing ability levels. The homogeneous group was comprised of students equal in number to and matched to the ability levels of the band and at-risk students, plus other students of differing ability levels matched to the ability levels of the additional students in the heterogeneous group. Data for a 6-month period were gathered to determine if a causal relationship existed between the groupings and student behaviors. The data consisted of direct observations of interactions between students and information obtained from school records, including: (1) number of days absent from school; (2) number of assignments to in-school suspensions; and (3) number of days of out-of-school suspensions. The school records showed that there were fewer absences, in-school suspensions, and out-of-school suspensions among the students who were heterogeneously grouped than homogeneously grouped. In addition, there were lower rates of undesirable student behavior and academic failure for the heterogeneous group than for the homogeneous group. (HOD)