This study examines experimentally the differences among four class sizes ranging from 16 to 37 pupils. Sixty-two grades 4 and 5 classes in 11 schools in Toronto participated in the study. The study investigated the effects of class size on teachers' expectations about the effects of specific class sizes; the attitudes and opinions of students, teachers, and parents; student achievement in reading, mathematics, composition, and art; the academic self-concept of students; and a variety of classroom process variables. The data indicate that varying class size in these two grades resulted in few changes in classroom functioning or in pupil achievement, except in acquisition of mathematics concepts. It is easy to conclude that reducing class size is perceived as beneficial by teachers. Students' attitudes toward school, however, are unlikely to change; neither are teachers' instructional styles. (Author)