Class composition and student achievement in elementary schools

Principals and teachers sometimes intentionally use class formation procedures to create class compositions. This study examined the class distributional properties of 200 elementary school classes in two school districts. Fifty-six classes were combination classes-classes with students from two grade levels. The two districts differed in their policies toward gifted and talented classes and in principals' preferences for heterogeneous student assignment. Students completed pretest measures of ability and independence and posttest measures of achievement. Hierarchical linear modeling procedures were used to estimate composition effects. The major findings were that (a) principals and teachers assigned higher ability and more independent students to combination classes; (b) these purposeful assignment procedures affected the class distributional properties and how achievement variation was allocated within and between classes; and (c) composition effects were observed for both ability and independence in the district with a commitment to gifted and talented classes, and these composition effects were strongest in combination classes. It is argued that evidence on both student assignment rules and substantive mechanisms for compositional effects are needed to properly interpret regression coefficients relating outcome to pretest in classroom studies.